HASA T#€litiicml M€m^mh4mn ^Sl0
<\
H
LM;17267E
NASA Technical Memorandum 89810
Reports of Planetary Geology
and Geophysics Program— 1986
NASA Ojftce of Space Science and Applications
Washington, D.C,
fUASA
National Aeronautics
and Space Administration
Scientific and Technical
Information Branch
1987
FOREWORD
This is a compilation of abstracts of reports from Principal
Investigators of NASA's Office of Space Science and Applications,
Solar System Exploration Division, Planetary Geology and Geophysics
Program.
The purpose of this publication is to document in summary form
research work conducted in this program over the past year (1986).
Each report reflects significant accomplishments within the area
of the author's funded grant or contract.
No attempt has been made to introduce editorial or stylistic
uniformity; on the contrary, the style of each report is that
of the Principal Investigator and may best portray his research.
Joseph Boyce
Discipline Scientist
Planetary Geoscience Program
CONTENTS
Foreword. iil
CHAPTER 1 - URANIAN SATELLITES AND RINGS
The Cratering Record at Uranus: Implications for
Satellite Evolution and the Origin of Impacting
Objects 3
R.G. Strom
Crater Morphology and Morphometry on the Uranian
Satellites 6
S.K. Croft
Cratering History of Miranda. 9
J.B. Plescia and J.M. Boyce
The Viscosity of Miranda 12
P.J. Thomas, R.T. Reynolds, S.W. Squyres, and
P.M. Cassen
Mechanical and Thermal Properties of Planetologically
Important Ices 15
S.K. Croft
Could Ariel Have Been Heated by Tidal Friction? 18
S.J. Peale
Geology and Cratering History of Ariel 19
J.B. Plescia and J.M. Boyce
On the Lack of Commensurabilities in the Mean Motions of
the Satellites of Uranus and the Resurfacing of Ariel 22
S.J. Peale
Orbital Dynamics of the Uranian Satellites Based on
Voyager Data 23
J.B. Plescia
Why No Orbital Resonances Among the Satellites of Uranus? 26
S.J. Peale
Voyager Observations of 1985U1 27
P. Thomas and J. Veverka
Kinematics and Dynamics of the Uranian Rings 30
R.G. French
CHAPTER 2 - OUTER PLANETS AND SATELLITES
Shapes of Small Satellites 35
P. Thomas
Accretional Heating of the Satellites of Saturn
and Uranus • 37
S.W. Squyres, R.T. Reynolds, A.L. Summers, and F. Shung
Investigations of the Surfaces and Interiors of
Outer Planet Satellites • 40
G. Shubert
Failure Strength of Icy Lithospheres 43
M.P. Golombek and W.B. Banerdt
Studies of Outer Planet Satellites, Mercury,
and Uranus • • 46
W.B. McKinnon and P.M. Schenk
S0„ on lo : A Thermodynamic Perspective 49
A. P. Zent and F.P. Fanale
Investigations of Planetary Ring Phenomena 52
J. A. Burns
Subcentimeter-Size Particle Distribution Functions in
Planetary Rings from Voyager Radio and Photopolarimeter
Occultation Data. 55
H.A. Zebker, G.L. Tyler, and E.A. Marouf
Observational Studies of Saturn's Rings. 57
C.C. Porco
The Production of "Braids" in Saturn's F Ring 59
J.J. Lissauer and S.J. Peale
Experimental Studies on the Impact Properties of
Water Ice 60
F.G. Bridges, D.N.C. Lin, and A. P. Hatzes
CHAPTER 3 - COMETS AND ASTEROIDS
Temperatures and Minimum Thickness of the Inactive
Surface Layer of Comet Halley 65
F.P. Fanale and J.R. Salvail
Comet Thermal Modeling 66
P.R. Weissman and H.H. Kieffer
The Loss and Depth of C0„ Ice in Comet Nuclei.
F.P. Fanale and J.R. Salvail
69
Dynamics of Long Period Comets 70
P.R. Weissman
VI
Observations on the Magnitude-Frequency Distribution
of Earth-Crossing Asteroids. 72
E.M. Shoemaker and C.S. Shoemaker
Asteroid Families, Dynamics, and Astrometry. 75
J.G. Williams and J. Gibson
1986 DA and 1986 EB: M-Class Asteroids in
Near-Earth Orbits 77
J. Gradie and E. Tedesco
Evolution of the Inner Asteroid Belt: Paradigms and
Paradoxes from Spectral Studies 80
M.J. Gaffey
Meteorite Spectroscopy and Characterization of
Asteroid Surface Materials 81
M.J. Gaffey
Source of the Optical Red-Slope in Iron-Rich Meteorites 84
CM. Pieters and P.H. Schultz
A Search for Spectral Alteration Effects in Chondritic
Gas-Rich Breccias 85
J.F. Bell and K. Keil
Analytical Electron Microscopy of Fine-Grained Phases in
Primitive Interplanetary Dust Particles and Carbonaceous
Chondrites 87
I.D.R. Mackinnon, F.J.M. Rietmeijer, and D.S. McKay
Asteroid Lightcurve Inversion . 90
S.J. Ostro and R. Connelly
Accumulation of the Planets 92
G.W. Wetherill
CHAPTER 4 - COSMOGONY AND DYNAMICS
Protostellar Disks and the Primitive Solar Nebula.... 97
P.M. Cassen, J.B. Pollack, T. Bunch, 0. Hubickyj,
P. Moins, and C. Yuan
Timescales for Planetary Accretion and the Structure of
the Protoplanetary Disk 100
J.J. Lissauer
Formation of Giant Molecular Clouds in Global Spiral ■
Structures: The Role of Orbital Dynamics and
Cloud-Cloud Collisions 101
W.W. Roberts, Jr. and G.R. Stewart
Analysis of Planetary Evolution with Emphasis on
Differentiation and Dynamics 102
W.M. Kaula and W.I. Newman
vii
Evolution of Planeteslmal Velocities 105
G.R. Stewart and G.W. Wetherill
Accumulation of Solid Bodies in the Solar Nebula 106
S.J. Weidenschilling and D.R. Davis
A scaling Law for Accretion Zone Sizes 108
Y. Greenzweig and J.J. Lissauer
Uranus and Neptune: Questions and Possible Answers 109
R.T. Reynolds and M. Podolak
Collisional and Dynamical Processes in Moon and
Planet Formation 112
C.R. Chapman, D.R. Davis, S.J. Weidenschilling,
W.K. Hartmann, and D. Spaute
Dynamical Constraints on the Origin of the Moon 115
A. P. Boss and S.J. Peale
Lunar Science from Lunar Laser Ranging 116
J.G. Williams, X.X. Newhall, and J.O. Dickey
Solar System Dynamics 117
J. Wisdom
Orbital Resonances, Unusual Configurations and Exotic
Rotation States Among Planetary Satellites 120
S.J. Peale
Dynamics of Satellites, Asteroids, and Rings 121
S.F. Dermott
Hamiltonian Theory of Nonlinear Waves in Planetary Rings.. 124
G.R. Stewart
Planetary Ring Dynamics and Morphology. 125
J.N. Cuzzi, R.H. Durisen, and F.H. Shu
Dynamical Studies of Saturn's Rings.... 128
P.D. Nicholson and C.C. Porco
CHAPTER 5 - PLANETARY INTERIORS, PETROLOGY, AND GEOCHEMISTRY
High Pressure Cosmochemistry of Major Planetary Interiors:
Laboratory Studies of the Water-Rich Region of the
System Ammonia-Water 133
M. Nicol, M. Johnson, S. Boone, and H. Cynn
Experiments Pertaining to the Formation and
Equilibration of Planetary Cores. .............................. 136
R. Jeanloz
VIII
Melting of Troilite at High Pressure in a Diamond
Cell by Laser Heating 138
W.A. Bassett and M.S. Weathers
Properties of Planetary Fluids at High Pressure and
Temperature 141
W.J. Nellis, D.C. Hamilton, N.C. Holmes,
H.B. Radousky, F.H. Ree . M. Ross, D.A. Young,
and M. Nicol
Gravity Data Analysis 144
W.L. Sjogren
The Origin of Polarity Asymmetries in the History of the
Geomagnetic Field. 147
E.H. Levy
Magnetic Flares in the Protoplanetary Nebula and the
Origin of Meteorite Chondrules 147
E.H. Levy and S. Araki
Generation of a Dynamo Magnetic Field in a Protoplanetary
Accretion Disk 148
T. Stepinski and E.H. Levy
The Steady State Toroidal Magnetic Field at the
Core-Mantle Boundary 148
S.J. Pearce and E.H. Levy
Lunar Magnetization Concentrations (Magcons) Antipodal
to Young Large Impact Basins 149
R.P. Lin, K.A. Anderson, and L.L. Hood
Implications of Convection in the Moon and the
Terrestrial Planets. 151
D.L. Turcotte
Global Petrologic Variations on the Moon: A Ternary-Diagram
Approach. 152
P. A. Davis and P.D. Spudis
A Chemical and Petrological Model of the Lunar Crust...... 155
P.D. Spudis and P. A. Davis
Calcium Carbonate and Calcium Sulfate in Martian
Meteorite EETA79001 158
J.L. Gooding and S.J. Wentworth
The Case for a Wet, Warm Climate on Early Mars............ 160
J.B. Pollack, J.F. Kasting, S.M. Richardson, and
K. Poliakoff
The Role of Regolith Adsorption in the Transition from
Early to Late Mars Climate 161
P.P. Fanale, S.E. Postawko, A. P. Zent, and J.R. Salvail
IX
CHAPTER 6 - SPECTROSCOPY AND REMOTE SENSING
Oxidized Basalts on the Surface of Venus:
Compositional Implications of Measured
Spectral Properties 165
CM. Pieters, W. Patterson, S. Pratt, J.W. Head, and
J. Garvin
Variations of Martian Surface Albedo: Evidence for
Yearly Dust Deposition and Removal 167
P.R. Christensen
On the Spectral Reflectance Properties of Materials
Exp-osed at the Viking Landing Sites 170
E. Guinness, R. Arvidson, M. Dale-Bannister, R. Singer,
and E. Bruckenthal
Spectral Mixture Modeling: Further Analysis of Rock and
Soil Types at the Viking Lander Sites... 173
J.B. Adams and M.O. Smith
Hisingerite and Iddingsite on Mars: Degradation of
Iron-Rich Basalts 175
R.G. Burns
Gossans on Mars: Spectral Features Attributed to Jarosite 176
R.G. Burns
Characterization of Surficial Units on Mars Using
Viking Orbiter Multispectral Image and Thermal Data 178
M.A. Presley, R.E. Arvidson, and P.R. Christensen
C0„ : Adsorption on Palagonite and the Martian Regolith. ....... . 181
A. P. Zent, F.P. Fanale, and S.E. Postawko
Investigation of Martian H„0 and C0„ Via Gamma-Ray
Spectroscopy • 184
S.W. Squyres and L.G. Evans
Reflectance Spectra of Mafic Silicates and Phyllosilicates
from .6 to 4.6 m. . 187
T.L. Roush, R.B. Singer, and T.B. McCord
Spectral Effects of Dehydration on Phyllosilicates........ 190
E.A. Bruckenthal and R.B. Singer
Studies of the Scattering/Absorption Properties
of Minerals 193
R.N. Clark
Mid-Infrared Spectroscopic Investigation. 196
J.W. Salisbury, L. Walter, and N. Vergo
Thermal-Infrared Spectral Observations of
Geologic Materials in Emission 199
P.R. Christensen and S.J. Luth
Bidirectional Reflectance Properties of Planetary
Surface Materials 202
B. Buratti, W. Smythe, R. Nelson, V. Gharakhani,
and B. Hapke
Atlas of Reflectance Spectra of Terrestrial, Lunar,
and Meteoritic Powders and Frosts from 92 to 1800 nm 205
J. Wagner, B. Hapke, and E. Wells
Deconvolution of Spectra for Intimate Mixtures 206
J.F. Mustard, CM. Pieters, and S.F. Pratt
Compositional Information for the Moon: Some Characteristics
of Current Near-IR Spectra (Telescopic and Laboratory)......... 208
CM. Pieters
Compositional Stratigraphy of Crustal Material from
Near-Infrared Spectra. 210
CM. Pieters
Preliminary Results of Spectral Reflectance Studies of
Tycho Crater 211
B.R. Hawke, CR. Coombs, P.G. Lucey, J.F. Bell,
R. Jaumann, G. Neukum, and CM. Pieters
Preliminary Results of Geologic and Remote Sensing
Studies of Rima Mozart 214
CR. Coombs and B.R. Hawke
Photometric Properties of Lunar Terrains Derived from
Hapke ' s Equation . 217
P. Helfenstein and J. Veverka
Bidirectional Reflectance Spectroscopy. 4. The Extinction
of Coefficient and the Opposition Effect..... 219
B. Hapke
On the Sputter Alteration of Regoliths of Outer Solar
System Bodies 220
B. Hapke
Charged Particle Modification of Surfaces in the Outer
Solar System. 221
R.E. Johnson
lo: Comparison of Photometric Scans Produced by the
Minnaert and Hapke Functions 223
D.P. Simonelli and J. Veverka
XI
Sulfur-Oxygen Processes on lo. . . 226
R.M. Nelson and W.D. Smythe
Solid Sulfur in Vacuum: Sublimation Effects on Surface
Microtexture, color, and Spectral Reflectance, and
Applications to Planetary Surfaces. 227
D.B. Nash
A Preliminary Analysis of the Mariner 10 Color
Ratio Map of Mercury 229
B. Rava and B. Hapke
CHAPTER 7 - RADAR APPLICATIONS
A Compilation System for Venus Radar Mission (Magellan)... 233
S.S.C. Wu, F.J. Schafer, and A.-E. Howington
New Very High Resolution Radar Studies of the Moon 237
P.J. Mouginis-Mark and B. Campbell
High Resolution Radar Map of the Moon. 238
T.W. Thompson
Landform Identification - Lunar Radar Images 240
H.J. Moore and T.W. Thompson
Goldstone Radar Observations of Mars: The 1986 Opposition. 243
T.W. Thompson
A Diffuse Radar Scattering Model from Martian Surface Rocks.... 245
W.M. Calvin, B.M. Jakosky, and P.R. Chrlstensen
Mars: Seasonally Variable Radar Reflectivity.... 248
L.E. Roth, R.S. Saunders, and T.W. Thompson
Measurements of the Dielectric Constants for
Planetary Volatiles 251
V.G. Anicich and W.T. Huntress, Jr.
Application of Numerical Methods to Planetary
Radiowave Scattering 253
R.A. Simpson and G.L. Tyler
CHAPTER 8 - AEOLIAN PROCESSES
Bright Sand/Dark Dust: The Identification of Active
Sand Surfaces on the Earth and Mars 257
H.G. Blount II, R. Greeley, P.R. Christensen, and
R. Arvidson
XII
Regional Sources and Sinks of Dust on Mars:
Viking Observations of Cerberus, Solis Planum, and
Syrtis Major. 259
S.W. Lee
High Resolution Thermal Infrared Mapping of
Martian Channels. . • 261
R.A. Craddock, R. Greeley, and P.R. Christensen
Mars: Morphology of Southern Hemisphere Intracrater
Dunefields. 264
N. Lancaster and R. Greeley
Timing of Frost Deposition on Martian Dunes: A Clue to
Properties of Dune Particles? 266
P. Thomas
Wind Ripples in Low Density Atmospheres. . . • 268
J.S. Miller, J.R. Marshall, and R. Greeley
Eolian Saltation on Mars. . • 271
D.J. MacKinnon
Mass Transport by Aeolian Saltation on Earth, Mars and
Venus: The Effects of Full Saltation Cloud Development
and Choking • • • 274
S.H. Williams and R. Greeley
Development of Wind Tunnel Techniques for the Solution
of Problems in Planetary Aeolian Processes 276
R. Sullivan, J. Lee, and R. Greeley
Aeolian Abrasion on Venus: Preliminary Results from
the Venus Simulator 279
J.R. Marshall, R. Greeley, D.W. Tucker, and
J.B. Pollack
Reynolds Number Effects on Surface Shear Stress Patterns
Around Isolated Hemispheres 282
J. A. Lee and R. Greeley
Determination of Surface Shear Stress with the
Naphthalene Sublimation Technique. . 285
J. A. Lee and R. Greeley
CHAPTER 9 - FLUVIAL PROCESSES
Groundwater Sapping Channels: Summary of Effects of
Experiments with Varied Stratigraphy. 291
R.C. Kochel and D.W. Simmons
Fluvial Valleys on Martian Volcanoes 294
V.R. Baker and V.C. Gulick
XIII
Valley Development on Hawaiian Volcanoes. . 297
V.R. Baker and V.C. Gulick
The Application of Flow Competence Evaluations to the
Assessment of Flood-Flow Velocities and Stresses... 300
P.D. Komar
Possible Origin of Some Channels on Alba Patera, Mars 303
S.E. Postawko and P. Mouginis-Mark
Non-Equilibrium Freezing of Water-ice in Sandy Basaltic
Regoliths and Implications for Fluidized Debris Flows on Mars.. 305
J.L. Gooding
Volatile Reservoirs Below the Surface of the Elysium
Region of Mars: Geomorphic Evidence 307
E.H. Christiansen and J. A. Hopler
Evidence for Glaciation in Elysium 310
D.M. Anderson
Water and Ice on Mars: Evidence from Valles Marineris 313
B.K. Lucchitta
Formation of the Layered Deposits in the Valles
Marineris, Mars 316
S.S. Nedell and S.W. Squyres
Geomorphic Evidence for Ancient Seas in West Deuteronilus
Mensae, Mars. 1. Regional Geomorphology. 319
T.J. Parker, D.M. Schneeberger , D.C. Fieri, and
R.S. Saunders
Geomorphic Evidence for Ancient Seas in West Deuteronilus
Mensae, Mars. 2. From Very High Resolution Viking Orbiter
Images 322
T.J. Parker, D.M. Schneeberger, D.C. Pieri, and
R.S. Saunders
CHAPTER 10 - VOLCANIC PROCESSES AND LANDFORMS
Explosive Volcanic Deposits on Mars: Preliminary
Investigations 327
D.A. Crown, L.A. Leshin, and R. Greeley
Pseudocraters As Indicators of Ground Ice on Mars.... 330
H. Frey
Eruptive History of the Elysium Volcanic Province of Mars 333
K.L. Tanaka and D.H. Scott
Late-Stage Flood Lavas in the Elysium Region, Mars............. 336
J.B. Plescia
XIV
Relative Ages of Lava Flows at Alba
Patera, Mars. 339
D.M. Schneeberger and D.C. Fieri
Martian Volcanism: Festoon-Like Ridges on Terrestrial
Basalt Flows and Implications for Mars 342
E. Theilig and R. Greeley
Studies of Fluid Instabilities in Flows of
Lava and D'bris 345
J.H. Fink
Long, Paired A'A/Pahoehoe Flows of Nauna Loa: Volcanoligical
Significance and Insights They Provide Into Volcano
Plumbing Systems , 348
S.K. Rowland and G.P.L. Walker
The 1984 Mauna Loa Eruption- and Planetary Geology 351
H.J. Moore
Calculated Viscosity-Distance Dependence for Some
Actively Flowing Lavas . 354
D. Pierl
Toward a Model for Leveed Lava Flows 357
S. Baloga
Studies of Vesicle Distribution Patterns in
Hawaiian Lavas. 360
G.P.L. Walker
Crustal and Subcrustal Nodules in Ejecta from
Kilbourne Hole Maar, New Mexico 362
J.L. Whitford-Stark
Geology of the Mohon Mountain Volcanic Field, Yavapai and
Mohave Counties, Arizona: A Preliminary Report 365
A.M. Simmons and J.S. King
The Geology of Pine and Crater Buttes: Two Basaltic
Constructs on the Far Eastern Snake River Plain. 368
P.F Mazierski and J.S. King
The Geology of Picacho Butte, A Silicic Volcanic Dome in
Northwest Arizona. 371
A. P. Kisiel and J.S. King
CHAPTER 11 - CRATERING PROCESSES AND CHRONOLOGIES
Computer Simulations of 10-km-Diameter Asteroid Impacts
into Oceanic and Continental Sites ............................ 377
D.J. Roddy, S.H. Schuster, M. Rosenblatt, L.B. Grant,
P.J. Hassig, and K.N. Kreyenhagen
XV
Impacts of Hemispherical Granular Targets:
Implications for Global Impacts 380
P.H. Schultz, D.E. Gault, and D. Crawford
Impact Vaporization: Late Time Phenomena from
Experiments • 382
P.H. Schultz and D.E. Gault
Oblique Impact: Projectile Richochet, Concomitant Ejecta,
and Momentum Transfer 384
D.E. Gault and P.H. Schultz
Momentum Transfer from Oblique Impacts 386
P.H. Schultz and D.E. Gault
Experimental Studies of Collision and Fragmentation
Phenomena • 388
W.K. Hartmann, D.R. Davis, and S.J. Weidenschilling
Centrifuge Impact Cratering Experiments: Scaling Laws for
Non-Porous Targets. 391
R.M. Schmidt
Impact Crater Scaling Laws 392
K.A. Holsapple
Experimental Evidence for Non-Proportional Growth of
Large Craters 394
P.H. Schultz and D.E. Gablt
Centrifuge Impact Cratering Experiments: Scaling Laws for
Non-Porous Targets 396
R.M. Schmidt
Impact and Cratering Processes on Asteroids, Satellites,
and Planets. 399
C.R. Chapman, D.R. Davis, and S.J. Weidenschilling
Crater Production on Venus and Earth by Asteroid and
Comet Impact 402
E.M. Shoemaker and R.F. Wolfe
The Surface Age of Venus: Applying the Terrestrial
Cratering Rate 405
G.G. Schaber, E.M. Shoemaker, and R.C. Kozak
A Statistical Study of Mercurian Crater Classes Applied
to the Emplacement of the Intercrater Plains 408
A. Woronow and K. Love
Studies of Early Intense Cratering and Possible
Saturation Effects. . . 411
W.K. Hartmann
Revision of the Martian Relative Age Chronology. 413
N.G. Barlow
XVI
Early Changes in Gradation Styles and Rates on Mars 416
P.H. Schultz and D. Britt
Cratering and Obliteration History of the South Polar
Region of Mars . . • • 418
J.J. Plaut, R.E. Arvidson, E.A. Guinness, and R. Kahn
Martian Rampart Craters: Morphologic Clues for the
Physical State of the Target at Time of Impact. 420
P.J. Mouginis-Mark
Crater Ejecta Morphology and the Presence of Water on Mars 423
P.H. Schultz
Thermal Inertia Characteristics of the Martian
Crater Curie. 426
V.M. Horner and J.R. Zimbelman
Dome Craters on Ganymede 429
J.M. Moore and M.C. Malin
Non-Newtonian Ice Rheology and the Retention of Craters
on Ganymede • • 432
P.J. Thomas and G. Schubert
Ejecta Types on Ganymede and Callisto 435
V.M. Horner and R. Greeley
Shatter Cones in Illinois: Evidence for Meteoritic Impacts
at Glasford and Des Plaines ........... 438
J.F. McHone, M.L. Sargent, and W.J. Nelson
Search for, and Study of. Paleozoic Impact Ejecta 439
W.F. Read
CHAPTER 12 - STRUCTURE, TECTONICS, AND STRATIGRAPHY
Comments on the Tectonism of Venus 443
R.C. Kozak and G.G. Schaber
Lithospheric Structure on Venus from Tectonic Modelling of
Compressional Features 446
W.B. Banerdt and M.P. Golombek
Venusian Tectonics: Convective Coupling to the Lithosphere? . . . . 449
R.J. Phillips
Bilateral Symmetry Across Aphrodite Terra. 452
L.S. Grumpier, J.W. Head, and D.B. Campbell
Secular Cooling of Earth as a Source of Intraplate Stress...... 455
S.C. Solomon
A Test of the Hypothesis That Impact-Induced Fractures
Are Preferred Sites for Later Tectonic Activity 458
S.C. Solomon and E.D. Duxbury
XVII
Physiographic Constraints on the Origin of Lunar
Wrinkle Ridges 461
M.P. Golombek and B.J. Franklin
Lithospheric Loading and Tectonics of the Lunar
Irregular Maria 464
J.L. Hall and S.C. Solomon
Characteristic Structures of the Highland Boundary on Mars:
Evidence Against a Single Mega-Impact Event?..... 467
A.M. Semeniuk and H. Frey
The Martian Crustal Dichotomy: Product of Accretion and
Not a Specific Event? 469
H. Frey, R.A. Schultz, and T.A. Maxwell
Implications of Viking Color Data for Evolution of the
Amenthes Region, Mars 472
T.A. Maxwell
Timing of Ancient Extensional Tectonic Features on Mars 474
R. Wichman and P.H. Schultz
Stress History of the Tharsis Region, Mars 476
R.A. Francis
History and Morphology of Faulting in the Noctis
Labyrinthus-Claritas Fossae Region of Mars 478
K.L. Tanaka and P. A. Davis
Thin and Thick-Skinned Deformation in the Tharsis
Region of Mars 481
T.R. Watters
Flexurally Resisted Uplift of the Tharsis Province, Mars 484
R.J. Phillips and N.H. Sleep
Tectonic Domains in the Eastern Hemisphere of Mars 487
T.A. Maxwell and C.E. Leff
Age of Fracturing and Mesa Development in the Elysium
Area, Northern Martian Plains 489
G.E. McGill
Geometry and Relative Age of Large Patterned Fractures in
Southern Acidalia Planitia, Mars. 492
M.C. Borrello
Origin of Fractures , Martian Polygonal Terrain. 495
L.S. Hills
The Effect of Overburden Thickness on Tension Fracture
Patterns Above an Uplifting Dome 497
H. Pranger II
Natural Fracture Systems on Planetary Surfaces: Genetic
Classification and Pattern Randomness 499
L.A. Rossbacher
xviii
Curvilinear Ridges and Related Features in Southwest
Cydonia Mensae , Mars 502
T.J. Parker, D.M. Schneeberger , D.C. Fieri, and
R.S. Saunders
Local-Scale Stratigraphy of Grooved Terrain on Ganymede....... 505
S.L. Murchie, J.W. Head, P. Helfenstein, and
J.B. Plescia
Stratigraphy of the South Polar Region of Ganymede... 508
R.A. De Hon
Creep of Ice: Further Studies. 511
H.C. Heard, W.B. Durham, and S.H. Kirby
Observations of Industrial Sulfur Flows and Implications
for lo 514
S.W. Lee, D.A. Crown, N. Lancaster, and R. Greeley
CHAPTER 13 - GEOLOGIC MAPPING, CARTOGRAPHY, AND GEODESY
Geology of lo 519
R. Greeley, R.A. Craddock, D.A. Crown, L.A. Leshin,
and G.G. Schaber
Large Scale Topography of lo. . . 522
R.W. Gaskell and S.P. Synnott
The Galilean Satellite Geological Mapping Program, 1986....... 524
B.K. Lucchitta
Cartography of Irregularly Shaped Satellites 525
R.M. Batson and K. Edwards
Voyager Cartography 527
R.M. Batson, E.M. Lee, and K.F. MuUins
Aspects of Voyager Photogrammetry . 528
S.S.C. WU, F.J. Schafer, R. Jordan, and A.E. Howington
Progress in Compilation of the 1 : 2 ,000,000-Scale
Topographic Map Series of Mars . 530
S.S.C. Wu, R. Jordan, and F.J. Schafer
Mars Digital Terrain Mode 531
S.S.C. Wu and A.E. Howington
Digital Cartography of Mars 534
R.M. Batson
The Control Network of Mars : October 1986 536
M.E. Davies
A Unified Lunar Control Network. 537
M.E. Davies, T.R. Colvin, and D.L. Meyer
XIX
The Control Networks of the Satellites of vJupiter. 539
M.E. Davies
CHAPTER 14 - SPECIAL Programs
Mlcrovax-Based Data Management and Reduction System
for the Regional Planetary Image Facilities 543
R. Arvidson, E. Guinness, S. Slavney, and B. Weiss
Martian Terrains 545
H. Masursky, M.G. Chapman, P. A. Davis,
A.L. Dial, Jr., and M.E. Strobell
The Mars Observer Camera 548
M.C. Malin, G.E. Danielson, A. P. Ingersoll,
H. Masursky, J. Veverka, T. Soulanille, and M. Ravine
Planetary Nomenclature 551
M.E. Strobell and H. Masursky
Enhanced Landsat Images of Antarctica and Planetary
Exploration 554
B.K. Lucchitta, J. A. Bowell, K. Edwards, E.M. Eliason,
and H.M. Ferguson
Author Index. 555
XX
CHAPTER 1
URANIAN SATELLITES AND RINGS
THE CRATERING RECORD AT URANUS s IBSPLICATIONS FOR SATELLITE
EVOLUTION AND THE ORIGIN OF IMPACTING OBJECTS,
Robert G, Strom, Department of Planetary Sciences,
University of Arizona^ Tucson, Arizona 85721
The crater size/ frequency distributions on the major
Uranian satellites show two distinctly different crater
populations of different ages (Smith et al. , 1986). Figure
1 is an "R" plot of the size/frequency distributions on
Oberon, Titania, Umbriel, Ariel, and the heavily cratered
and the resurfaced regions of Miranda. Oberon, Umbriel, and
the heavily cratered terrain on Miranda have the same lunar-
like size distribution. The crater size/frequency
distributions on Titania, Ariel, and the resurfaced areas of
Miranda are quite different. They are characterized by an
overabundance of small craters and a paucity of large
craters relative to Oberon, Umbriel, and the heavily
cratered terrain on Miranda, At diameters greater than 30
km, the crater density is significantly less on Titania than
Oberon and Umbriel, and progressively decreases from Titania
to Ariel to the resurfaced areas of Miranda. Furthermore,
the paucity of large craters and a corresponding abundance
of small craters becomes more pronounced with decreasing
crater density, i.e, with time. This suggests that the
objects which caused this younger crater population evolved
with time by mutual collision where the collision of large
objects produced more and more small ones. If so, they must
have been in planetocentric orbits for frequent collisions
to have occurred. Since only the young crater population
occurs on Titania, the older crater population must have
been largely obliterated by a resurfacing event.
Although the young crater population is only recognized
on relatively young surfaces, it must also be present on the
older heavily cratered surfaces as well. Thus, the old
crater population is a mixture of the young population and
an original old population that can be recovered by
subtracting Titania' s young population. This "unmodified"
population (curve Ul in Figure 2a) is quite different from
other crater populations in the Solar System (see Figure
2a).
Any hypothesis on the origin of the objects responsible
for the period of heavy bombardment must account for the
occurrence of different crater populations (size /frequency
distributions) in different parts of the Solar System, One
hypothesis suggests that an early high flux of comets was
responsible for the period of heavy bombardment throughout
the Solar System (Shoemaker and Wolfe, 19'82), To test this
hypothesis, a computer simulation using short-period comet
impact velocities and a modified Holsapple-Schmidt crater
scaling law was used to recover the size distribution of
coraetary nuclei from the observed crater ing record. Figure
2b shows the results of this simulation. It shows that if
comets on short-period-like orbits were responsible for the
crater ing record ^ they must have had radically different
size distributions in different parts of the Solar System.
In fact^ their diameters would have to have been on average
larger in the inner Solar System where comets rapidly lose
mass than at Jupiter where their masses are conserved » This
is highly unlikely, and suggests that comets were not
responsible for the period of heavy bombardment.
The most likely explanation for the cratering record is
that the period of heavy bombardment was caused by different
families of accretional remnants indigenous to the system in
which the different crater populations occur. Since the
same crater population is found on all the terrestrial
planets but not Jupiter^ this family of accretional remnants
was indigenous to the inner Solar System and confined to
heliocentric orbits with small semi -major axes (< 3 AU) .
The satellites of Jupiter, Saturn, and Uranus each have
different crater populations suggesting that they were the
result of accretional remnants in planetocentric orbits
around each of these planets.
Since the young crater population at Uranus shows
evidence that it was formed by objects that evolved by
mutual collisions in planetocentric orbits, it is quite
possible that both crater populations resulted from one
family of accretional remnants. The old crater population
has a paucity of small craters which is what one would
expect from the accretional process where large objects are
built from smaller ones. If these objects had their
relative velocities increased by close encounters with the
satellites, then they could collide, resulting in a
progressive depletion of large objects and a corresponding
increase in small objects as characterizes the younger
crater population. In this case, at least the initial
orbits would have low eccentricities and the impact rate
would be about the same on all major satellites. Therefore,
differences in crater density would represent the relative
age of surfaces among the satellites and can be used to date
the relative time of resurfacing events. Since the crater
density on Oberon and Umbriel is significantly lower than on
Miranda, it suggests that both Oberon and Umbriel were
resurfaced near the end accretion, and that the sequence of
resurfacing events from oldest to youngest was, 1) Oberon
and Umbriel, 2) Titania, 3) Ariel, and 4) Miranda.
References ;
Smith, et al, , Science , 233 , p. 43, 1985.
Shoemaker, E.M. and Wolfe, R.F., Satellites of Jupiter ,
Univ. of Arizona Press, p. 277, 1982.
-2
■d— j-j-jkAAtJ
10 100 1000
CROTER OWiETER Ital
I® 100 lOM
CiOTEi DIAMETER Ctal
Figure 1. "R" plot of the crater size/ frequency
distributions on Oberon (O) , Umbriel (U), Titania
(T), Ariel (A), the heavily cratered terrain
(MHC) and resurfaced terrain (MLC) on Miranda,
and the lunar highlands (LH)»
K» too l@©0
I to m mi
PROJECTILE WaiETER ItaJ
Figure 2. (a) The crater size/frequency distributions
representing the period of heavy bombardment on
the terrestrial planets (Tl), at Jupiter (Jl), at
Saturn (SI) and at Uranus (Ul). (b) The
projectile size-frequency distributions for
short-period comets recovered from the cratering
record.
ClATER MORPHOLOGY AUD IK5RPHQMETRY OS THE OHANIMf SATELLITES,
Steven K, Croft , Lunar and Planetary Laboratory, University
of Arizona, Tucson, Arizona 85721,
Crater Morphology » Fresh craters on the icy Uranian
satellites exhibit a range of morphologies similar to
craters on the icy satellites of Jupiter and Saturn.
Craters on the large satellites Ganymede and Callisto show a
progression of crater morphologies with increasing crater
diameter; simple craters, complex craters, pit craters, and
multiring structures at the largest diameters. Craters on
the smaller Saturnian satellites are almost exclusively
simple or complex in morphology, pit craters being generally
absent. Craters on the Uranian satellites also include
simple and complex craters, but no obvious multiring basins
occur. However, the two largest craters on Titania, with
diameters of about 335 Jon (about lO'S lat. , 280° long.) and
146 km (about -15°, 40°), appear to be pit craters. Complex
craters on the Uranian satellites show the same general
structural features found in complex craters on the other
icy satellites? a) central structures are usually single
peaks, circular or linear elongate in planform, A large
conical peak approximately 8 km high visible on Oberon's
limb is morphologically similar to the distinctive large
central peak in Izanagi on Rhea. Albedo patterns and
irregularities in Oberon's limb around the peak suggest a
crater rim around 300 km in diameter, again comparable to
Izanagi. b) Terrace structures are very underdeveloped
compared to terraces in complex craters on the terrestrial
planets. No true terraces are apparent, and only slump
deposits at the bases of rim scarps are found (though this
may be a resolution effect). c) The planforms of complex
crater rims are frequently polygonal like complex craters on
rocky planets.
Simple -Complex Transition Diameters . Estimates of
transition diameters from simple to complex crater
morphologies are given for the five large Uranian satellites
and 1985U1 in Table 1, and plotted with transition diameters
on other bodies against surface gravity in figure 1, The
transitions on Ariel and Titania are well determined. The
other transitions are only bracketed by approximate limits.
The lower limit on Miranda is well determined, but the upper
limit is set by a single degraded structure that is complex
if it is a single crater and not a degraded cluster. The
upper limits on Umbriel and Oberon are well defined, but the
lower limits are somewhat uncertain due to poor resolution.
The lower limit on 1985U1 is based on a single large crater
which (again subject to resolution limits) appears to be
simple. Two simple theories of crater modification (1,2)
suggest an inverse correlation of the transition diameter
with surface gravity. The difference in transition diameter
between sedimentary (s) and crystalline (c) rocks on the
Earth suggest a direct dependence on target strength or
density as well (1 , 3 ) ^ a dependence further indicated by the
general downward shift of the transition diameters on the
icy satellites relative to the rocky terrestrial planets.
However^ adopting a g"^ dependence for the icy satellites
seems to indicate two ice sequences; one including
Ganymede, Callisto and most of the Uranian satellites, and a
second including the Saturnian satellites, Miranda, and
Ariel, One might suggest a compositional difference was
responsible since the Uranian and Jovian satellites have
similar uncompressed bulk-densities while the Saturnian
satellites are somewhat less dense, but Titania and Ariel
with the best - determined transition diameters are on
separate "sequences," Apparently other factors affect
transition diameters not fully accounted for in current
theories .
Possible Large-scale Impacts . One of the unusual dark
structures on Miranda, the "banded ovoid" (1) centered near
IS'S lat. and 50° long,, occupies a quasi-circular area
about 320 km in diameter (figure 1). Crater counts on the
dark material and the surrounding light "highlands"
demonstrate that the dark terrain is substantially younger
than the light terrain. Stereo imagery and limb profiles
show that the light terrain slopes down to the contact with
the dark terrain, suggesting that the dark terrain fills a
previously existing depression. This old depression is
inferred to be an impact scar by several observations s 1)
strings of overlapping craters morphologically similar to
secondary crater chains around large lunar basins extend
radially away from the depression center along the
terminator. The largest individual craters in the string
are 14 to 18 km in diameter, appropriate for a 300 km
primary crater ( 4 ) . 2 ) Several fractures and two valleys
similar to Vallis Bouvard at Orientale Basin on the Moon
extend radially from the depression. 3) A tongue of
material similar to ejecta deposits around lunar basins
obliterates the rim of a 60 km crater ("A") adjacent to the
depression and covers most of the floor. 4) Craters on the
light terrain fall into two preservation classes; fresh and
sharp rimmed, and extremely subdued and degraded with
virtually no intermediate states. In addition, a layer of
bright material at least 1 km thick appears to cover most of
the visible hemisphere (5). Both of these observations may
be accounted for if the depression is an impact crater; the
light layer representing the thick ejecta blanket that
covered and "softened" all pre-existing topography, thus
providing a stratigraphic horizon upon which later fresh
appearing craters formed.
Another possible modified large-scale impact is located
near ICS lat. and 30° long, on Ariel (figure 2). The
structure is a roughly circular depression about 245 km in
diameter surrounded by massifs that occur in linear chains
that radiate away from the depression center and merge into
the surrounding dissected plateau region. An excess of 10-
12 km diameter craters that are the right size and radial
distance for a 200-250 km diameter primary crater occur on
plateaus surrounding the depression. Several large valleys
are also sub-radial to the depression. The floors of the
depression and the radial valleys are entirely covered by
the smooth material that covers the floors of the large
linear valleys to the south and west. The general
arrangement is reminiscent of some of the large impact
basins on Mars^ such as Argyre and Chryse, which act as
topographic lows towards which flow has occurred.
References
1, Croft J, S.K^ (1982) Saturn; Program and Abstracts ^ p.
77 » 2, Melosh, H.J. (1982) J. Geophys. Res, 87, p. 371-
380e 3« Pike, R«J. (1980) PLPSC 11th, p. 2159-2189. 4.
Croft, S.K« (1986) submitted. 5. Smithy B.A,^ et al.
(1986) Science 233, p, 43-64«
TABLE 1, SIMPLE-CC^LEX CRATER «»MSITIONS ON TEE URANIAM SATELLITES
Name
Radius*^ km Density*, g/cra^ Surface
Simple-Complex
Gravity cm/s^
Transition Diara.
,km.
1985U1
85
±
5
(1.4)
(3.3)
>48
Miranda
242
±
5
1,26 ± 0.39
8.5 ± 2.6
« 28 -^ 41
Ariel
580
±
5
1.65 ± 0.30
26.7 ± 5.1
13 ± 2
Urabriel
595
±
10
1.44 ± 0»28
23.9 ± 5,1
= 23 ^ 38
Titania
805
±
5
1,59 ± 0.09
35.8 ± 2.2
24 ± 2
Oberon
775
±
10
1.50 ± 0.10
32.5 ± 2.6
~ 29 ^ 45
*Radii and densities adapted from Smith et al. (1986) except for
density of 1985 Ul which is a nomimal average uncompressed
density for other satellites.
Ariel
Figure 2.
Figure 1.
ClATSRire fflSTOSf OF MIKAIDA J.B. Plescia, U.S.G.S., Flagstaff, AZ 86001, and
J.M. Boyce, NASA Headquarters, Washington, D.C. 20546
The surface of the southern hemisphere of Miranda Imaged by Voyager 2
(Smith et al. , 1986) is divisible into two general types of terrain: cratered
terrain, characterized by numerous craters and undulating intercrater plains;
and basins, circular to rectangular areas of complex morphology having large-
scale albedo markings. (The term "basin" is used without a genetic
implication.) To determine the relative ages of the terrains and the length
of geologic activity, crater-frequency data were compiled for various parts of
the cratered terrain and the basins (Table 1).
All of the well-resolved impact craters are either bowl shaped or flat
floored; craters having central peaks or rings were not observed. Almost all
craters >20 km in diameter are degraded and some are barely detectable.
Smaller craters range in appearance from fresh to degraded. Ejecta deposits
are generally absent, although two craters are surrounded by low-albedo ejecta
material.
The cratered terrain has a varied crater frequency that, in part, is
related to position on the satellite. From the anti-Uranian hemisphere (long
180 ) to the sub-Uranian hemisphere (long ), from areas A to B to C, crater
frequencies decrease by an order of magnitude (Table 1). Differences in
crater frequencies and the shape of the cumulative size-frequency
distributions indicate a resurfacing of the cratered terrain, the sub-Uranian
hemisphere having been more affected than the anti-Uranian hemisphere.
Within the cratered terrain in the anti-Uranian hemisphere, near lat
-50 , long 180 , the size-frequency distributions correspond to those of
steady-state surfaces on the Moon. This correspondence indicates that the
cratered terrain is locally saturated with craters <7 km in diameter.
A gradient in the cratering rate from the leading to the trailing
hemisphere (Shoemaker and Wolfe, 1982; Horedt and Neukum, 1984) can explain
the observed pattern of resurfacing and the correlation of crater frequencies
with position within the cratered terrain. The data suggest that Miranda
originally was oriented about 90 from its present orientation, such that the
most heavily cratered areas that are now near long 180 were originally near
the apex of motion (long 90 ). Resurfacing within the cratered terrain
resulted from the ballistic distribution of ejecta that eroded and burled all
craters to some extent but was most effective on the small ones. Apparently,
as a result of the formation of the basins, Miranda's moments of inertia were
changed, resulting in a reorientation of the satellite.
The banded basin in the leading hemisphere consists of a smooth interior
area and a marginal area characterized by albedo banding resulting from
concentric, outward-facing scarps. The interior has a higher crater frequency
than the margin (Table 1). Although the frequency difference is statistically
significant, it may not reflect a real difference In formation ages: the
extensive faulting in the margin might have obliterated craters, thereby
lowering the observed crater frequency and the apparent age.
The grooved basin in the trailing hemisphere consists of an interior
area, characterized by randomly oriented ridges and troughs, and a marginal
area, characterized by parallel, closely spaced ridges and grooves. The
interior has a significantly higher crater frequency than the margin (Table
1). All of the craters within the margin are superposed; none appear to be
affected by the grooving, indicating that the frequency represents a formation
age rather than a crater-retention age.
The polar basin was divided into four areas: the bright chevron-shaped
feature, grooved area A (near lat -70 , long 300 ), grooved area B (near lat
-55°, long 340°), and a boundary area that partly surrounds the other areas.
Three areas — the chevron, grooved area A, and the boundary area — have similar
crater frequencies (Table 1), indicating that they developed
contemporaneously. Grooved area B has statistically higher crater
frequencies, which may represent a crater-retention rather than a formation
age. This area may have formed by the tectonic disruption of adjacent
cratered terrain and it may have inherited some preexisting craters. This may
explain why the crater frequencies for grooved area B are lower than those
characteristic of the cratered terrain but higher than those of a "new"
surface of the same age that was originally uncratered.
Crater-frequency data indicate that the cratered terrain is the oldest
terrain on Miranda and that it has been locally resurfaced. Ages of the
grooved and banded basins are similar but younger than the cratered terrain,
and the polar basin is the youngest terrain. Observed crater frequencies on
Miranda are, in general, lower than those on any of the other Uranian
satellites, indicating that Miranda's surfaces are the youngest in the
system. Miranda's relative youth is even more apparent when the increase of
the cratering rate inward from Oberon is considered; the cratering rate at
Miranda is about 14 times that at Oberon (Smith et al., 1986). Thus, a
surface on Miranda dating to the same period as does Oberon's would have about
20,250 craters >^ 10 km in diameter/lO km . However, the observed crater
frequencies on Miranda (Table 1) are several orders of magnitude lower.
Apparently, the recorded geologic history of Miranda began after the heavy
bombardment had ended and after much of the activity on the other satellites
had ceased.
BEFERENCESs Horedt, G.P., and Neukum, G., 1984, Icarus, 6^, 710-717;
Shoemaker, E.M. , and Wolfe, R. , 1982, Satellites of Jupiter, 277-339; Smith,
B. A., et al., 1986, Science, 233, 43-64,
10
TABLE 1
MIRMBA TERRMM CRATER FRSJOENCIES
FREQUENCY OF CRA.TSRS > B / 10^ KM^
DIAlffiTER (KM)
REGION
I
2
5
10
20
POLAR BASIN
Grooved A
10763+1345
252^651
(283+218)
(58+ 99)
(12+45)
Grooved B
22038+1805
5916+935
908+366
(227+183)
(60f94)
Chevron
16440+1911
2222+703
(156+186)
(21+ 68)
(2+21)
Boundary
8561+1468
1740+662
348+292
(91+151)
(21+73)
Average
15157+ 846
3352+398
338+126
(56+51)
(10+22)
GROOVED BASIN
Margin
(7263+559)
1737+273
355+124
151+ 81
(16+26)
Interior
(10325+1042)
3371+596
794+289
206+147
(77+90)
BANDED BASIN
Margin
(1E5+2206)
(12115+ 768)
7304-188
115+ 75
18+29
Interior
(93886+2057)
(59948+1644)
1307+243
232+102
60+51
CRATERED TERRAIN
A (-70°/235°)
15599+1789
7800+1265
21754-668
1017+457
(415+292)
B (-40°/270°)
11034+1077
3153+576
736+278
155+126
(44+68)
C (-30°/315°)
724O+-800
1126+315
581+123
435+196
177+125
D (-65°/130°)
38312+1880
9045+914
1477+369
233+146
(62+76)
E (-75°/175°)
(30037+3440)
9846+1969
2757+1042
(924+603)
(325+358)
F (-65°/175°)
(32483+1250)
10419+708
2357+337
(667+179)
(235+106)
G (-35°/195°)
(32650fl472)
10325+828
2523+409
598+199
(25^130)
H (-45°/225°)
(25230fl627)
7658+896
2046+463
5904-249
(195+143)
Numbers in parentheses are extrapolations of the data to diameters at which
craters of a given size were not present (at the large-diameter end) or at
which resolution, viewing, or lighting limitations prohibited accurate counting
(at the small-diameter end).
11
Th® Viscosity of Miranda
P.J. Thomas, R.T. Reyaolds, S.W. Squyres and P.M. Cassen
Theoretical Studies Branch, NASA Ames Research Center, Moffett Field, CA 94035.
Voyager 2 images of Miranda reveal a significant history of geological activity. Overlyiag an apparently
ancient, cratered terrain are assemblages of concentric ridges, scarps and dcirk banded material. Three
regions of complex teirain are visible in the Voyager images, which cover virtually an entire hemisphere.
These regions are typically ~ 200A:m in diameter.
Although the concentric ridges have some similarity to the regularity of the grooved terrain on
Ganymede, they axe also associated with what appeeir to be extensive flow regions (possibly associated
with internal melting) that, in some cases, have modified the geometry of the ridges. Finally, widespread
brittle behavior is observed in the form of large faults, scarps and graben up to 15km deep.
We examine the problems that evolutionjiry thermal and structural models of Miranda must
face, to provide an convincing explanation for such topographic complexity. The problems center around
the requirement that such a small body [R ~ 242km) must have a sufficient heat generation mechanism
to lower its viscosity substantially. It would generally be expected that a body of such size with such
low eunbient temperatures (^ 50 K) would have an extremely rigid surface. While the observed fault
regions are consistent with a rigid surface, the high degree of sphericity of Miranda implies a low viscosity
material, at least at some time in Miranda's past.
For viscous relaxation on a scale of L ~ 100A;m to occur over a reasonable timescale (r, ^ 10^ y),
the dominant viscosity jjs. determining viscous motion is given by
KTrLpi
'*~ 4jr ^ '
where for Miranda g = O.OQms"^ and p = 1200kgm~^ (Tyler et al., 1986) yielding pi^lO'^'^Paa. As-
suming a Newtonian viscosity law for ice (Weertman, 1970):
jji = lio exp
(^*(^-l)) (2)
with parameters appropriate to low-temperature ice A* — 25; Tm = 273K; and ^o = 10^*Pas (Reynolds
and Cassen, 1979), this is appropriate to a temperature T of 157ir, or a temperature > lOOK above the
ambient temperature.
The two reference models for Miranda that we consider are that (1) the complex terrain was
emplaced during a period of reaccretion following a disruptive impact (Smith et al., 1986) in this case
the energy required for resurfacing is produced by the release of the potential energy of tne reaccreted
fragments; or that (2) the complex terrain represents a disturbed region overlying a diapiric instability,
similar to a mechanism proposed for the formation of palimpsests on Ganymede (Hale et al., 1980).
In the second case the energy required for resurfacing must have an internal origin, such as radiogenic
heating or tidal deformation. These two models represent exogenic and endogenic formation models
for the complex terrain, respectively. They also permit estimates of dynamic viscosities and associated
emergy inputs required by these models.
]f population statistics derived for the sizes of comet nuclei are valid for projectiles impacting on
Miranda, it is conceivable that Miranda underwent a series of collisions with bodies of sufficient kinetic
energy to disrupt it in the first Gy after its initial formation (Shoemaker and Wolfe, 1982; Smith et al.,
1986). This model implies that the final stages of reaccretion may have produced the observed terrain
patterns by liberation of the gravitational potential energy of the disrupted fragments of the satellite.
While both models must be able to account for Miranda's high degree of sphericity, the reac-
cretion model faces a more stringent relaxation condition: that relaxation must remove all traces of the
impact crater associated with the disruptive impctct. Such a crater will have topographic harmonics
associated with the rim with wavelengths much smaller than 100km and thus much longer relaxation
times. For example, to relax crater rims with wavelengths on a scale of 10km. in lO^y would require a
viscosity of lO^^Paa.
Unfortunately for this model, the energy that can be obtained from reaccretional heating for
a body as small as Miranda is limited. The thermal energy retained by Miranda cannot be significantly
greater than the binding energy, -^ 2 X 10^ Jkg~^ which, given a specific heat for ice of ~ 10^ Jkg~^K~^,
12
predicts a maximum global temperature increase as small as ~ 20K. This estimate is in good agreement
with more detailed models (Squyres et al., 1986). Of course, local heating arising from point contacts
between reaccreted fragments may much greater temperatures in small regions. Furthermore, disruption
and reaccretion will dissipate much of this initial energy, tending to produce a more isothermal temper-
ature profile (Ellsworth and Schubert, 1983). One final important problem that this mechanism must
face is that, according to vaj-ious evolutionary models of the icy satellites, radiogenic heating produces
a maximum temperature profile for small satellites within the first ~ lOOMy (Ellsworth and Schubert,
1983; Federico and Lanciano, 1983). This heat will also be dissipated by a cycle of multiple disruption
and accretion, if such exists.
We can consider energy constraints on the endogenic model by calculating the ascent time Ta
for a buoyant inviscid sphere (our diapir model) of radius a = 50km at low Reynolds number through a
layer of thickness L = 100A;m:
Apga^
we assume a density difference Ap of 300A;grm~^, appropriate to warm pure ice surrounded by material
with the mean density of Miranda. Assuming Ta^lO^y we obtain /z^ 10^^ Fas, consistent with the
determination above. This value is, however, very much of an upper limit, because Miranda's surface
gravitational acceleration was assumed for g. The cooling time for such a diapir is o^/k ~ 3 X lO^y
(where k ~ 10~^trPs~^). For this ascent time (more reasonable than lO^y because of the presence of
some cratering on the ridged terrain) we have n^lO^^ Pas whichjis appropriate, using the viscosity law
above, to T = 16QK, or ~ lOOif above Miranda's ambient temperature.
It is possible that extinct radionuclides such as ^^Al may have had a significant energy input
into Miranda in the past. However, the half life of such elements (typically ^10®j/) indicates that any
resurfacing that occurred as a result would be substantially older than the observed, very lightly cratered
terrain on parts of Miranda.
It appears to be clear that some energy source other than accretional heating is required to
provide both the spherical shape of Miranda and the evidence of substantial viscous flow activity. The
exogenic model requires enhanced temperatures after the last reaccretion, placing serious constaints on
short term heating processes. In addition, it has a serious problem in accounting for the absence of
small-scale impact features on much of Miranda's surface.
Given the difficulties in finding a suitable energy source capable of raising Miranda's global
temperature to a sufficient extent to reduce the viscosity of pure ice adequately, it is possible that
mobilization of ice clathrated with CO, iVj or CII4 by pressure-solution creep may be an important
transport process. Under certain conditions, viscosities many orders of magnitude smaller than those
calculated above may be produced by this mechanism (Stevenson and Lunine, 1986).
Since the temperatures determined here lie above the H2O.NH3 eutectic at 173if, internal
melting may have occurred to some extent. If the flow features were emplaced significantly after the
accretional period (^10*y) then tidal heating alone remains as a plausible energy source. The energy
required to produce increases in temperature of the extent calculated for the entire satellite is of the order
of 10^* J. If this was expended over the period of 3 x lO^y mentioned above, a heating rate of ~ lO^W,
which is ~ 10~* the heating rate associated with lo's tidal flexure. For Miranda no easily indenfiable
orbital resonance can account for this enhanced heating (Squyres et al., 1985). However, the effects of
mutual satellite pertubations have not yet been considered in full complexity (Dermott and Nicholson,
1986). In conclusion, it should be emphasized that due to the inability of reaccretional heating to provide
a important energy input for Miranda, any heating mechanism required by the endogenic model must
also be required by the exogenic model.
References
Dermott, S.F. and F.D. Nicholson (1986). Masses of the satellites of Saturn. Nature, S19, 115-120.
Ellsworth, K. and G. Schubert (1983). Saturn's icy satellites: thermal and structural models. Icarus, 54,
490-510.
Federico, C. and P. Lanciano (1983). Thermal and structural evolution of four satellites of Saturn.
Annales Geophysicae, 1, 469-476.
13
Hale, W., J.W. Head and E.M. Parmentier (1980). Origin of the Valhalla ring structure: alternative
models (abstract), in. Papers presented to the Conference on Multi-Ring Basins: Formation and
Evolution, pp. 30-32, Lunax and Planetary Institute, Houston.
Reynolds, R.T. and P.M. Cassen (1979). On the internal structure of the major satellites of the outer
planets. Geophys. Res. Lett., 6, 121-124.
Shoemaker, E.M. and R.F. Wolfe (1982). Gratering time scales for the Galilean satellites. In Satellites
of Jupiter D. Morrison, ed.). University of Arizona Press, Tucson.
Smith, B.A. and the Voyager Imaging Team (1986). Voyager 2 in the Uranian System: Imaging Science
Results, Science, 2SS, 43-64.
Squyres, S.W., R.T. Reynolds and J.J. Lissauer (1985). The enigma of the Uranian satellites' orbital
eccentricities. Icarus, 61, 218-223.
Squyres, S.W., R.T. Reynolds, A.L. Summers and F. Shung (1986). Accretional Heating of the Satellites
of Saturn and Uranus. This volume.
Stevenson, D.J. and J.I. Lunine (1986). Mobilization of cryogenic ice in outer Solar System, satellites.
Nature, S2S, 46-48.
Tyler, G.L. and the Voyager Radio Science Team (1986). Voyager 2 radio science observations of the
Uranian system: atmosphere, rings and satellites, Science, 28S, 79-84.
Weertman, J. (1970). The creep strength of the Earth's mantle. Rev. Geophys. Space Phys., 8, 145-168.
14
MECTMICM. MS) THESIttL mOFlETIES OF PMMETOMGIGMXY IMFOITAIT ICES Steven
K. Crofts Lunar and Planetary Laboratory s University of Arizona s Tucson s
Arizona 85721
Two "sequences" of ice composition have been proposed for the icy
satellites I 1) a dense nebula model (1) in which C and N are CH^ and NH3J
yielding^ in order of decreasing condensation temperature, hydrated rockj
H2O ice, NH3'H20j CH^-ice clathrate and finally solid CH^i and 2) a solar
nebula model (2) in which C and N are in CO and N2 yielding hydrated rock,
H2O ice and finally CO- or N2-ice clathrate. Careful modeling of the
structure^ compositions and thermal history of satellites composed of these
various ices requires quantitative information on the; Ddensity,
2)compressibilityj and 3)thermal expansion of the rock and ices
(incorporated as an equation of state or EOS), as well as the 4)heat
capacity and 5) thermal conductivity. EOS's and thermal data have been
given previously for the H2O ices (3) and CH^ (4). However j much recent
interest in structures of molecular ices by physical chemists has resulted
in a large increase in the data available on ices of planetological
interest. Thusj data have been gathered from the literature to update
equations of material properties for the H2O and CH^ ices and to calculate
such equations for the first time for NH39 CO, N2 and CO2 ices.
Pure Ice Data . EOS's have been fit to the density data of the molecular
ices shown in table 1. The # phases includes the nvimber of solid phases
plus one liquid phase (except for CH^'5.75H20) analyzed for each compound.
Compared to silicate minerals, the bulk moduli (K^q) of the ices are one to
two orders of magnitude smaller, the K' 's are about the same, and the
K'-j's are about half. The inferred thermal expansions show enormous
variability, increasing from rock to H2O ice to the other molecular ices.
The effect is not trivial s increasing the interior temperature of an icy
satellite from accretion temperatures near 100°K to around BOCK yields a
0.2% volume increase in the rocks, ~ 3% increase in the H2O ices, and > 30%
increase in the other molecular ices. The assembled heat capacity data
show the same trends, increasing dramatically from rock to H2O ice to the
other ices. Of particular note here is that the heat capacity of rock
assiimed in previous models was near 1.2 Joules/g°K, appropriate for the
high interior temperatures of rocky planets, but far too large for the lOOK
to 300K temperatures appropriate to the icy satellites. The net affect on
icy satellite thermal histories, because of the large mass fraction of the
rock, is to shorten the times for internal heating and melting by factors
of 2x to 3x. Finally, the assembled thermal conductivity data (figure 1)
again show a definite trends decreasing from rock to H2O ice to the other
molecular ices. Of particular interest is the extremely low thermal
conductivity of CH^-clathrate compared to H2O ice or rock - a point
returned to below.
Binary/Multiple Ice Systems . The simple molecular ices represent end-
nxombers of more complex ice "minerals" that form when the simple ices are
mixed. Ammonia hydrates (NH3-nH20) and ice clathrates (CH/^, N29 C0'nH20)
are likely forms (5,6). The thermodynamics of ice clathrates have been
studied by (7). A preliminary EOS for CH^- and N2-clathrates have been
constructed here from the meager data available, but much work needs to be
done to directly determine thermal expansion and bulk modulus detivatives.
Data for the NH3-H20 system are limiteds heat capacities and latent heats,
liquid densities near room temperature, single density measurements
2NH3-H20, NH3=H20 and NH3"2H20, and some high pressure measurements at
15
selected compositions (8). Approximate bounds on the thermal and
mechanical properties of the ammonia hydrates may be set by looking at the
properties of the end-members? NH3 and H2O. For example j heat -capacity
data for the ammonia hydrates are intermediate between ammonia and water.
However » actual measurements are preferred and thus analytical,
experimental s and theoretical studies of the ammonia hydrates have been
initiated. An equation for the density of liquid NH3-H2O (at all NH3
concentrations) has been constructed using previous high temperature data
and our own low temperature data. The liquid densities have been combined
with thermodynamic and solid density data to determine the low pressure
portion of the NH3 -21120 phase diagram. Combining our results with the high
pressure data, we infer at least two solid phases of NH3 -21120 with the
transition near Ikb and with a fairly large density difference. A phase
change at such a low pressure is an important one structurally even for
fairly small icy satellites.
Applications . Detailed structural and thermal modeling of icy satellites
has commenced based on the new data. For example, one of the real
surprises of the Voyager Uranus encounter was the presence of endogenic
resurfacing and fracturing on tiny Miranda. The implied volume increase of
Miranda represented by the fractures is ~ 6%. How does Miranda become warm
enough to exhibit such activity? One possible approach is illustrated by
the thermal profiles in figure 2. If Miranda consists of simple ice and
rock, the maximiun interval temperature rise is ~ 10°K, perhaps yielding
temperatures in the core capable of melting a little CH^ (if any is present
in pure form). Such small amounts would likely freeze en route to the
surface. However, if the ice is in the form of CH^ clathrate, the
temperature rise in the core is « 120°K, sufficient to melt CH^ through
much of the interior and probably to drive it through the resulting thin
crust to the surface (the dark flows on Miranda may be radiation-darkened
CH^ (9)), Even partial melting of eutectic NH3-H2O is possible (depending
on boundary conditions). The large internal heating due to clathrate and
the presence of molecular ices of high thermal expansivity may also account
for the observed fracturing via simple heating and thermal expansion.
Thus, the difference between resurfaced Miranda and primitive Mimas may
simply be the presence of clathrate. The effects of the clathrate vs. pure
ice on larger satellites is not as great (cf. Titania curves in fig, 2)
because convection provides a fairly low temperature cutoff (detailed
values depend on surface conditions and assvmiptions concerning viscosity).
Thus, the unusual thermal and mechanical properties of the molecular and
binary ices suggest a larger range of phenomena than previously
anticipated, sufficiently complex perhaps to account for many of the
unusual geologic phenomena foxind on the icy satellites.
References
1. Prinn R.G. and Fegley B, (1981) Ap.J. 249, 308-317. 2. Lewis J.S. and
Prinn R.G. (1980) Ap.J. 238 , 357-364. 3. Lupo M.J. and Lewis J.S. (1979)
Icarus 40, 157-170. 4. Lupo M.J, and Lewis J.S. (1980) Icarus 42, 29-34.
5. Miller S.L. (1985) in Ices in the Solar System , NATO, p. 59-79, 6.
Lewis J.S. (1972) Icarus 16 , 241-252. 7. Lunine J.I. and Stevenson D.J,
(1985) Ap. J. Supp. 58, 493-531. 8. Johnson M.L. et al^ (1985) in Ices in
the Solar System , NATO, p. 39-47. 9. Smith B.A. et al. (1986) Science
233, 43-64.
16
TAILS 1. MJLECIIMR ICffi
Compound //Phases
K,
oo
K'
K'
H20
8
NH3
4
CH4
4
^2
6
CO
5
CO2
3
CH4-5
75H2O
1
114 Kb
5.5
-0.101 Kb/'K
75
5.5
-0.186
21
4.9
-0.074
28
3.8
-0.108
20
6.6
-0.092
77
7.0
-0.183
=88
(5.5)
(-0.1)
T— i — r-n-TTT
10 ■mfaxraa.'i
, t^ \» -J-.< I » . -
f ,, —^ -,pw~y
:i
CiathTBte i»lts
■ ^ . — ' WH5-H20
Eutectic Mlts
^ I H2, CO atlt
^^., X I .m.^
O.S Et/Radius
Figure 1,
Figure 2.
17
Could Ariel have been He ated bv Tidal Friction?
S. J. Peak (University of California at Santa Barbara)
For significant dissipation in Ariel, one must hypothesize the past
existence of an orbital resonance to force and maintain a substantial
eccentricity in the orbit. The certain capture of Ariet/Umbriel into a
simple eccentricity resonance at the 2/1 commensurabllity of orbital
mean motions had they passed through singles out this strong reso-
nance as the most likely to have had any effect on Ariel's interior. An
Immediately apparent problem with the occupancy of this resonance
is the neccessity for an escape by an unknown mechanism. A lack of
resolution of this problem may mean that tidal expansion of the satel-
lite orbits has been insufficient for the 2/1 commensurabllity to have
been encountered, but the striking modifications of Ariel's surface
suggests that something much more dramatic has happened. A max-
imum dissipation occurs when Ariel's eccentricity has been forced to
an equilibrium value where the ratio of orbital mean motions of the
resonant pair is fixed as the orbits of the satellites continue to ex-
pand from tides raised on Uranus. The dissipation is then simply the
difference between the rate of work done by the tidal torques from
Uranus and the rate of orbital energy increase, where nearly all of
the dissipation occurs in Ariel. With dissipation dE/dt = 0,26nr
[n — mean motion, T = torque), we can infer a maximum efiecron
Ariel's thermal history by assuming the longest possible existence of
the resonance in equilibrium with the minimum average value of Q
for Uranus. Q^^m — ^^^^ (^2 — -104) is now determined by Miranda
starting at the "standstill" orbit. The maximum dissipation rate In
Ariel would have been 7 x 10^^ ergs/sec which would drop to 2.3 x 10^'
ergs/sec at the last position at which the 2/1 resonance could have
existed 2.6 x 10® years later — the resonance being disrupted at least
2 X 10® years ago. The absolute maximum dissipation rate would
be reduced both by the time necessary to establish the resonance in
the first place and by a consequently larger minimum Q. Subsolidus
convection may be able to remove the heat even at the highest rate
of deposition without disrupting a rigid lithosphere, but such a con-
clusion depends on highly uncertain parameters. The actual heating
rate has undoubtedly been much less than this maximum, so even the
former (unlikely?) existence of a 2/1 resonance may not be sufficient
to account for Ariel's smoothed surface. Miranda could never have
been trapped in any first order resonance, so Its even younger surface
is more a mystery.
18
GEOLOGY MD CRAXEMMG HISTORY OF ARIEL J.B. Plescias U.S.G.S., 2255 N. Gemini
Drive, Flagstaff, AZ 86OOI5 and J.M. Boyce, NASA Heaquarters, Washington, D.C.
20546
The surface of Ariel imaged by Voyager 2 (Smith et al»j 1986) can be
divided into several types of terrain on the basis of morphologyj cratered
terrain, subdued terrain, ridged terrain, and plains. Cratered terrain, at
lat -20°, long 0° and at lat -30°, long 255°, is characterized by a rolling
surface, east- or northeast-trending grabens and scarps, and scattered
superposed impact craters. Between the two occurrences of cratered terrain
are areas of subdued terrain characterized by highly degraded craters and
narrow (<3 km wide) ridges. Ridged terrain bounds the subdued terrain and
occurs as narrow bands within it. It is characterized by bands 25 to 70 km
wide within which are parallel, east- or northeast-trending ridges and troughs
typically about 10 to 35 km apart. Individual ridges and troughs extend 100
to 200 km, but the bands of ridged terrain extend many hundreds of
kilometers. At several locations, bands of ridged terrain are continuations
of well-defined grabens and follow the same structural trends. Plains fill
topographically low areas such as graben floors and irregular depressions.
The plains that partly fill the grabens locally exhibit a medial trough from
which plains-forming material appears to have flowed out onto the surface.
Crater statistics were compiled for each of the terrain types (Table
I). Despite differing morphology, the various terrains on Ariel do not
exhibit large variations in crater frequency. The difference in crater
frequency between the cratered terrain, the most heavily cratered surface, and
the plains, the least cratered surface, is only a factor of 3 to 4, in
distinct contrast to the order-of-magnitude differences observed on Miranda
(Plescia and Boyce, 1986; Smith et al., 1986).
Crater saturation of the cratered terrain occurs at diameters of about 12
km, whereas that for the subdued and ridged terrains is closer to 7 km,
suggesting that the cratered terrain is the oldest of the three. The ridged
terrain cuts both the subdued and cratered terrains and is therefore younger
than both. The plains, which are not saturated with craters, are the youngest
terrain and apparently formed at different periods of time, as indicated by
the crater-frequency data.
None of the observed surfaces on Ariel record the period of accretion.
If the surface of Ariel were as old as that of Oberon or Umbriel, whose
surfaces presumably reflect the period of accretion, it would have a frequency
of about 1,800 craters >_ 30 km/IO^ km^. However, the observed crater
frequencies on Ariel are clearly lower (Table 1). The presence of few craters
larger than 50 km in diameter also indicates that its observed surface was not
formed during accretion. Ariel appears to have been completely resurfaced
since it formed.
The extensive network of grabens on Ariel indicates that it has
experienced global tensional stresses. Freezing of an initially liquid water
interior and the resultant satellite-wide expansion may have produced the
necessary global tension. Because the grabens are locally floored by plains,
they must have formed, at least in part, during that portion of Ariel's
history when new terrains were being formed and when conditions were
appropriate for the mobilization of material for resurfacing.
Ariel's geologic history may be better understood by a comparison with
that of Dione, which has similar size, density, and surface temperature and
for which extensive thermal modeling has been done (Stevenson, 1982| Ellsworth
and Schubert, 1983). On the basis of the thermal models for Dione, solid-
19
state convection within Ariel can be estimated to have lasted for several
billion years, a situation conducive to the deformation of the brittle surface
layer. Temperatures in excess of the ammonia-water (NH-'HoO) eutectic melting
point ( 173 K) might be sustained for the first few hundred million years near
the surface and for perhaps 2 billion years at deeper levels. If clathrate
dissociation occurred catastrophically (Stevenson, 1982), then resurfacing
could have been produced by the explosive eruption of material onto the
surface. Because clathrate dissociation at shallow depths (<10 km) can occur
at temperatures lower than the melting points of either water ice or ammonia-
water ice mixtures, such resurfacing events would have been possible for a
longer periods of time than the estimates given above. Thus, it seems that
conditions appropriate for resurfacing could have occurred during Ariel ' s
early history.
BKFERKNCESs Ellsworth, K. , and Schubert, G. , 1983, Icarus, 54 , 490-510;
Plescia, J.B, and Boyce, J.M. , 1986, this volume; Smith, B.A. , et al.., 1986,
Science, 233, 43-64; Stevenson, D.J,, 1982, Nature, 298, 142-144.
20
miLE- 1
MIEL TEEXAIH GRATEE FREQBEMCIES
REGION
CDUNTIHG
ARM KIT
MEQUEWCT OP CIATEIK _> D / 10® KM^
DIAMETER (KM)
10
20
LAKEST DIST»
30 ClATER (KM) SLOPE
Cratered terrain
(lat -20°, long 0°)
Cratered terrain
(lat -30°, long 255°)
Subdued terrain
(combined areas)
Ridged terrain
to (combined areas)
Plains
(lat -45°, long 320°)
Plains
(lat -15°, long 30°)
Plains
(lat -15°, long 345°)
66962
37893
(22967+586) 2285+185 358+73
37684 (63908+1302) 4007+326 849+150
75+33 (19+17) 29
69+43 (34+30) 29
87706 (29663+582) 2497+169
376+66
62+27
46944 (37106+889) 2641+237 547+108
(11365^548) 1311+186 198+72
87+48
16819 (13769+905) 1873+33
725+208
33+19
69
23+22 (18+20) 20
10740 (24484+1510) 1955+427 162+123 (46+65) (15+37) 13
34+30 33
106+79 (68+64) 24
2.6
2.7
2.7
2 = 7
2.7
2.5
2.0
Numbers In parentheses indicate extrapolations of the data. At small diameters extrapolations are necessary due to
resolution limits, at large diameters because of a lack of craters.
On the Lack of Commensurabilities In the Mean Motions of the Satel-
lites of Uranus and the Resurfacing of Ariel
S. J. Peale (University of Califorina at Santa Barbara)
The lack of commensurabilities among the mean motions of the
satellites of Uranus is investigated by determining the probabilities of
capture into those orbital resonances which might have been encoun-
tered as the satellite orbits expand differentially from tidal torques.
For the minimum value of the dissipation function of Uranus Qu f^
6600, the Miranda/ Ariel pair would have passed through the 4/3, 3/2
and 5/3 commensurabilities but with no chance of capture into any
of the orbital resonances and consequently no resulting constraints on
the orbital evolution. Passage of Miranda through the inclination-
type resonances in the wrong direction for capture is unlikely to ac-
count for the observed 4° inclination. In contrast the Ariel/Umbriel
pair would have encountered the 2/1 commensurability with certain
capture into either one or both of two eccentricity-type orbital reso-
nances at 2/1 unless e^ > 0.027 and eu > 0.023 as the resonances are
approached. It is extremely unlikely that this pair could have avoided
being trapped into these resonances if, in fact, they had been encoun-
tered. Certain capture of Ariel/Umbriel into eccentricity- type reso-
nances at the 5/3 commensurability would require e^ < 0.0040 and
ej/ < .0047, so these resonances could have most likely been traversed.
Certain capture of Miranda/Urnbriel into the 3/1 eccentricity reso-
nances involving Miranda's parameters would occur if e^ < 0.0045 for
the simple e-type and eju < 0.0084 for the mixed e-type, which would
imply trapping Miranda/ Umbrlel Into this resonance if it were encoun-
tered. A possible history for the satellite system would be to avoid
traversal of those resonances into which capture was certain. With the
potential Love number &2C/ = 0.104 for Uranus, Qu ^ 11,000 would
mean that Ariel/Umbriel did not encounter the 2/1 commensurabil-
ity, Qu ^ 39,000 avoids the 3/1 Miranda/Umbriel commensurability,
and Qu ^ 100,000 starts the Arlel/Urabriel outside the 6/3 com-
mensurability leading to virtually no tidal expansion of the orbits at
all. An alternative to this uninteresting history Is investigated
to evaluate the conditions necessary to allow sufficient tidal heating
of Ariel for the observed resurfacing. The maximum tidal dissipaton
in Ariel (assumed to be locked into the 2/1 eccentricity-type reso-
nance with Umbrlel) is determined as a function of Ariel's separation
from Uranus. The tidal dissipation for the configuration correspond-
ing to the last possible existence of the resonance is only comparable
to a lunar like radiogenic source from the satellite core, but could be
about 30 times larger than this value in the earliest past. Countless
scenarios can be constructed which could lead to the resurfacing of
Ariel in the 2/1 resonance, although the constraints may make them
somewhat Implausible. In any case, detailed thermal histories of Ariel
are not warranted unless and until a supportable means of removing
Arlel/Umbriei from the 2/1 resonance is found.
22
ORBITAL DYNAMICS OF THE UEANIAl SATELLITES BASED OM VOYAGER DATA J.B»
Plesciaj U.S. Geological Surveys 2255 N. Gemini DrivCj Flagstaff, AZ 86001
The satellites of Uranus all have significant non-zero eccentricities and
in the case of Miranda a significant inclination as well (Table 1). Squyres
et al. (1985) studied the time scales and possible energy effects associated
with the orbital evolution of these bodies, specifically the eccentricities.
However, because of the uncertainty in the sizes and masses (data were based
on estimates of Brown et al., 1982 and Brown and Cruikshank, 1983), there was
a corresponding uncertainty in the results. The Voyager 2 encounter with
Uranus in January 1986- provided more accurate estimates of the masses and
sizes of the satellites (Smith et al., 1986) « As a result, it seems useful to
reexamine the orbital history and possible tidal heating by using the Voyager
data. This is particularly relevant because Miranda, which had the largest
pre-Voyager size uncertainty, is also the satellite which appears to have
undergone the most extensive endogenic evolution.
Miranda and, to a certain extent, Ariel have undergone significant
endogenic activity. Crater frequency data for Ariel (Plescia and Boyce,
1986a), in combination with the cratering time scale postulated by Smith et
al. (1986), indicate that Ariel's surface is >3.5 Gy old and that no major
resurfacing has occurred since. Miranda, on the other hand, has the least
cratered surface of any satellite in the Uranian system (Plescia and Boyce,
1986b) and may have surfaces as young as several hundred million years old.
The relevant orbital dynamics equations (Peale, 1977| Peale et al., 1980)
follow.
The change in orbital eccentricity (e ) with time (t) is
de/dt == - (n k^n M r^ e)/(2 m a^ Q) (1)
where ^^2 is the satellite potential Love number, n is the satellite mean
motion, m and M are the masses of the satellite and Uranus, r is the satellite
radius, a is the satellite semimajor axis and Q is the satellite dissipation
function. The equation is an approximation because the term relating to the
effect of dissipation within Uranus has not been included. Dissipation within
Uranus is negligible and dissipation within the satellite is the dominant
term.
The potential Love number (fe„ ) is
"■2
= (1.5)/[1 -h (19 u / 2 p g r)J (2)
where u is the satellite rigidity, p is the satellite density, and g the
surface gravity.
Finally, the homogeneous tidal heating rate (£") for a synchronously
rotating body is approximated as
dE/dt = [(m^ n^ v) / (u Q)] [0,396 e^ + 0.0566 o^] (3)
where o is the spin obliquity: the angle between the spin vector and the orbit
normal. If the satellite is in Cassini spin state 1 (c less than the orbital
inclination), there is no obliquity-decay heating. If, however, this is not
23
the case J the obliquity term can become important.
Assuming u = 4 X 10 dyne cm"^j a value typical for low- temperature ice,
and Q = 20; the orbital eccentricity decay times and homogeneous tidal heating
rates were calculated (Table 1)« These results are similar to those of
Squyres et al» (1985) and indicate that the calculated rates are relatively
insensitive to the uncertainties in mass and density of the satellites. The
calculations indicate that the present eccentricities of Miranda and Ariel
should decay on time scales of 10' years, whereas those of Umbriel, Titania^
and Oberon are stable for 10^ to 10^ years- The original eccentricity (e^)
iSj however, unknown and could have been quite large.
To make the present orbital eccentricity of Ariel and Miranda long lived,
such that a late-stage event to pump up the eccentricity is not neccessary,
requires that either u ox Q he increased. Neither parameter is well
constrained because the interior thermal conditions and compositions of the
satellites are uncertain* A § of >100 would produce an eccentricity decay
time of 10°-10° years for Ariel and Miranda, Similarly, increasing u would
also lengthen the decay time« Rigidities of 4 X 10^^ dvne cm~^, similar to
rock rather than ice, would increase the decay time to 10°-10' years.
When obliquity heating is not Important, the tidal heating rates from
eccentricity decay are about one to several orders of magnitude below that of
the radioactive heating (Table 1). The low tidal heating and low radioactive
heating of Ariel, Umbriel, Tltania, and Oberon are consistent with their
generally ancient age. Miranda, however, presents a significant problem
because of the youthful appearance of its surface.
The inclusion of the obliquity term in equation (3) has important
implications for Miranda, If Miranda occupies Cassinl state 2 (o greater than
the orbital inclination), then the obliquity term in equation (3) becomes
important- Assuming that all of the satellites have obliquities equal to
their inclinations, the tidal energy produced (JJ) was calculated (Table 1).
For Ariel, Umbriel, Titania, and Oberon, the term remains unimportant because
the obliquity is small. However, inclusion of the terra for Miranda Increases
the heating rate by 3 orders of magnitude*
Alternatively, if there is no obliquity heating of Miranda, the
eccentricity heating rate can only be raised to approximately that of
radioactivity. Enchanced eccentricity heating will result, relative to the
model In Table 1, by reducing either Q ox u^ For Q = I and u = 4X10-'-^ dyne
cm~^, the heating rate is 3,03 X 10^ erg sec" , corresponding to a heat flow
of 0«45 erss cm"^ sec" ■'"a For Q = 20 and u = 4 X 10^ dyne cm~2, the rate is
1,52 X 10^^ erg sec" , corresponding to a heat flow of 0,20 erg cm"'^ sec" *
lEFERENCESs Brown, R.H,, and Cruikshank, D,P., 1983, The Uranlan satellites;
Surface compositions and opposition brightness surges: Icarus, v. 55, p. 83-
92| Brown, R^H^, Cruikshank, D»P., and Morrison, D., 1982, Diameters and
albedos of satellites of Uranus; Nature, v, 300, p« 423-425; Peale, S.J.,
1977, Rotation histories of the natural satellites, _lri^ (Burns, J. A., ed,)
Planetary Satellites, p« 87-112; Peale, S.J., Cassen, P., and Reynolds, R.T,,
1980, Tidal dissipation, orbital evolution, and the nature of Saturn's inner
satellites; Icarus, v« 43, p. 65-72; Plescia, J»B., and Boyce, J.M. , 1986a,
Crater frequencies on Ariel, this volume; Plescia, J.B», and Boyce, J,M.,
1986b, Miranda cratering history, this volume; Smith, B,A, , Soderblom, L^A* ,
et al., 1986, Voyager 2 in the Uranlan system? Imaging science results:
24
Science, v. 233^ p. 43~64; SquyreSj S.W., Reynolds, R.T., and Lissauer, J.J.,
1985, The enigma of the Uranian satellites' orbital eccentricities; Icarus, v.
61, p. 218-223.
Table 1
Miranda
Ariel
Dabriel
Tltamia
Ofeeron
Radius (km)
Density (gm cm"-^)
Mass (gm)
Gravity (cm sec"^)
Semimajor axis (km)
Eccentricity
Inclination °
242+5
1,26+0.
7.48 X
8.52
129783
0.0027
4.2
.39
102^
580+5
1.65+0.
1.35 X
26.74
191239
0.0034
0.3
,3
1024
595+10
1.44+0.
1,27 X
23,94
265969
0.0050
0.4
,28
1024
805+5
1.5T+0.
3.47 X
35.76
435844
0.0022
0.1
.09
102^
775+10
1.50+0,10
2.92 X 102^^
32.48
582596
0.0008
0.1
de/dt iQ^ZO) sec"^
4.79 X
107
1.38 X
107
1.22 X
10^
8.16 X
10^
6,64 X 10^
tidal heating
erg sec"^ (Q=20)
J
II
1.52 X
1.61 X
loJA
10^'
1.02 X
1.36 X
10^6
1.70 X
2.17 X
10^5
10^5
8.18 X
8.91 X
10^3
10^3
8.37 X 10^^
1.41 X 10^2
tidal energy
heat flow
-2 -1
erg cm sec
I
II
0,02
21.88
0.24
0.32
0.04
0,05
0.001
0.001
0.00001
0.00002
radioactive
heat flow
-2 -1
erg cm ^ sec
0.15
0.93
0.64
1,13
0,91
J does not include the obliquity term in equation (3)
II includes the obliquity term in equation (3).
25
V^hy No Orbital Resonances Among the Satellites
<?f'Mra nus?
S. J. Peale (U. C. Santa Barbara)
Most of the orbital resonances among the satellites of the major
planets are thought to have been assembled by differential tidal ex-
pansion of their orbits. We can investigate why this has not occurred
for the Uranus system by determining the resonances which would
have been encountered for various values of Uranus' tidal effective
Q. If the minimum Q = 17,000 is assumed for Uranus (from Ariel's
proximity to Uranus), Miranda would have passed through the 4:3
and 3:2 orbital resonances with Ariel. As Ariel's orbit expands faster
than Miranda's, capture into the Miranda- Ariel resonances is impossi-
ble. Althrough the passage of Miranda through the resonances in the
wrong direction for possible capture would have increased Miranda's
orbital eccentricity and inclination, the latter increase cannot explain
the current 4° inclination for Miranda's orbit since the corresponding
6:4 and 8:6 inclination resonance passages would have increased the
inclination only a few tenths of a degree. For the same minimum Q,
Ariel would have passed through the 2:1 and 5:3 orbital resonances
with Urabriel. But Ariel and Umbriel would have been captured into
either of the lowest order 2:1 eccentricity resonances with certainty
unless gyi > 0.0224 and ej/ > 0.0219 at the time of encounter. As the
time constant for damping Ariel's eccentricity from tidal dissipation
within the satellite is only about 5 x 10^ years (with Q = 100 and
rigidity // = 4 x 10^^ assumed), and Ariel's current eccentricity being
only 0.0034, it Is very unlikely that Ariel could have avoided capture
into the 2:1 libration if the sj'stem in fact had passed through this
resonance. The Q of Uranus need only be increased to 30,000 to keep
Ariel always outside the 2:1 resonance with Umbriel. To avoid cer-
tain capture of Arlel-Umbriel in the 5:3 resonance, e^ > 0.0036 and
ey > 0.0059 at the time of passage. With current eccentricities at
0.0034 and 0.005 respectively, and the fact that eccentricities are re-
duced from initial values upon a probalistic escape from the resonance,
it is possible that the system passed through the 5:3 resonance. But if
the tidal Q of Uranus is larger than 200,000, Ariel would have started
even outside the 5:3 resonance. Since the Q of Jupiter is in the range
6 X 10^ to 2 X 10^, 200,000 is not an outrageous value of Q for Uranus.
This comparison and the possible escape of Ariel-Umbriel from the 5:3
resonance, means that it is not unreasonable that we find no orbital
resonances among these satellites.
26
VOYAGER OBSER¥ATIOiS OF 1985U1. P, Thomas and J» Veverka, Cornell
University.
Intrwiuctlon. Of the 10 small Uranian satellites^ 1985U1 is the largest
and the only one for which a resolved image was obtained by Voyager 2. In
terms of albedo^ the other nine satellites seem to be similar to 1985U1
(Smith et al, ^ 1985). Thus the single image of 1985U1 is important in sug-
gesting what these other objects may be like »
Size and Shape» Figure 1 shows the limb and terminator outline of the
satellite viewed from 61° latitude, 78° longitude. The limb of the satel-
lite can be measured to subpixel precision by techniques reported by
Dermott and Thomas (1986). Scans across the limb can be used to locate the
edge of the disk to about 0,3 pixels. The resultant line and sample coor-
dinates are scaled to kilometers using the range and camera focal length
(Davies and Katayama, 1981). The center was located approximately from the
position of the terminator and from the phase angle. Direct measurement of
the long and short dimensions gives 82 km and 77 km, respectively; an
ellipse fit to the limb using the same center coordinates gives axes of 78
and 76 km. Thus the satellite's average radius is somewhat less than 77
km^ and, at least in the orbital plane, the satellite is not greatly
elongated. The dimension perpendicular to the orbital plane is unknown.
Surface Features, One crater, approximately 45 km across, is near the
terminator. Two other possible craters are visible, but there is no basis
for attempting crater counts, or making any statements concerning crater
morphology.
Photometry, We have used a Minnaert function to approximate the photo-
metric behavior of the surface of 1985U1. Such an approximation works well
on other Uranian satellites for which extensive data were obtained. Addi-
tionally, in the case of 1985U1, the irregular shape of the satellite
introduces more serious uncertainties into any photometric analysis than do
details of the photometric function employed. We sampled the disk by two
methods, one a grid search of every other pixel, the other interactively
selected 2^2 pixel boxes. At each point we assumed that
I/F = Bq (a) iJ-o M- "
where -nf cos i is incident solar flux, I is scattered intensity, (iq is the
cosine of the incidence angle (i), ij, is cosine of the emission angle (e), k
is the Minnaert limb darkening parameter, and a is the phase angle. Assum-
ing an ellipsoidal shape allows a least squares fit from the Minnaert
parameters of Bq and k. The two sampling methods give nearly identical
results:
k = 0.65 ± 0.06 Bo = 0.043 ± 0.002
These values appear reasonable. For comparison with the Moon at a =
33°, Bo = 0,058 and k = 0.60 (Helfenstein and Veverka, 1986), while both
Titania and Oberon have k - 0.7 at this phase angle.
27
To obtain an estimate of the satellite's geometric. albedo and normal
reflectance we must extrapolate the Bq value determined at 33° to a = 0°.
Considering the overall similarity between the physical properties of
1985U1 and those of Phoebe, a reasonable guess for p-j would be 0,024 mag/
deg, the average value found for Phoebe by Thomas et al . (1983). The
resulting geometric albedo would be 0.09, with my = 20»5. Of the five
larger satellites of Uranus, Voyager phase data are most complete for
Titania. Analysis of disk-resolved photometry shows that for Titania Bq
increases by a factor of 1.63 in going from a = 33° to a = 0°, correspond-
ing to an effective value of Pi = 0.016 mag/deg. Since Titania is con-
siderably brighter than 1985111, this value is likely to be a lower limit to
Pi for 1985U1, Thus, p > 0.069 and my < +20.8, consistent with the
estimate made above.
Discussion^ In size and shape, as well as in albedo, 1985U1 seems to be
generally similar to Saturn's satellite Phoebe. (The average radius of
Phoebe is about 110 km.) Although we lack spectral data on 1985U1, if this
satellite resembles the other satellites of Uranus for which data are
available (Smith et al . , 1986), then it has an almost neutral gray color
very similar to that of Phoebe. The importance of the 1985U1 image is that
it provides the best tie-in of the normal reflectance of any of the newly
discovered satellites with that of the five larger ones and with that of
the rings. In this respect 1985U1 is considerably darker than the darkest
of the larger satellites (Umbriel) and is a little brighter than the cur-
rently accepted albedo for the ring particles (about 0.05, according to
Smith et al , , 1986).
This research was supported by NASA>grants NSG 7156 and NA6W-111.
REFERENCES
Davies, M. E., and F. Y. Katayama (1981). Coordinates of features on the
Galilean satellites. J. Geophys. Res . 86 (AlO) , 8635-8657.
Dermott, S. F., and P. C. Thomas (1986). ^The shape and internal structure
of Mimas, For submission to Icarus .
Helfenstein, P., and J. Veverka (1986). Photometric properties of lunar
terrains derived from Hapke's equation. For submission to Icarus .
Smith, B. A,, L. A. Soderblom, R, Beebe, D. Bliss, J, M. Boyce, A. Brahic,
G. A. Briggs, R. H, Brown, S, A. Collins, A. F. Cook II, S. K. Croft,
J. N. Cuzzi, G. E. Danielson, M, E. Davies, T. E. Dowling, D. Godfrey,
C. J. Hansen, C. Harris, G. E. Hunt, A. P. Ingersoll, T. V. Johnson,
R, J. Krauss, H, Masursky, D, Morrison, T. Owen, J. B.Plescia, J. B.
Pollack, C. C. Porco, K. Rages, C. Sagan, E. M. Shoemaker, L. A.
Sromovsky, C. Stoker, R. 6. Strom, V. E. Suomi , S. P.Synnott, R. J.
Terrile, P. Thomas, W. R. Thompson, J. Veverka (1986). Voyager 2 in
the Uranian system: Imaging science results. Science 233 , 43-64,
Thomas, P., J. Veverka, D. Morrison, M. Davies, and T. V, Johnson (1983.
Phoebe: Voyager 2 observations. J. Geophys. Res . 88, 8736-8742.
Thomas, P., J. Veverka, and S. Dermott (1986), Small Satellites, In
Satellites (J. Burns and M. Matthews, eds.), U. of Arizona Press,
Tucson, pp. 802-835.
28
too
50-
1
-50
-100
"T — f t — r i 1
— 1 — 1 — r- rT-| 1 -1 -I — r~ 1 i i ■ i |
^sM^*^-*^-^***^ 1985 Ul
— p-
f*
^ ^^^
/
^-,
-.
d
«
t
\
- /
t
\
i
f
i>
-
- 1
-
- {
—
- \
I
\
*T
: \
/
-
\
—
_
N »•'
-
-
^^%-^.--.'
-
1 1 1 1 1
1 1 1 ...J- I 1 ,,L_ 1 \ 1 1 \ J \ \ L
„_J_.
-100
-50
km
50
100
Figure. Outline of limb and terminator of 1985U1, Terminator points
connected by dashed lines.
1985U1 Summary
Orbit;
a = 86000 ± 50 km
P = 18.3 hrs
Spin Period:
Size/Shape:
Photometry:
Probably synchronous (18,3 hrs)
Apparent radii in image: 82 x 77 km
Ellipse fit: 78 x 76 km
Average radius: < 77 km
Minnaert parameters rBo = 0.043 ± 0.002
at a = 33° ^k = 0.65, ± 0.06
Estimated geometric albedo 0,07-0.09
Estimated opposition magnitude my ~ +20,5-21
29
Kinematics and Dynamics of the Uranian Rings
Richard G. French
Dept. of Earth, Atmospheric, and Planetary Sciences
Massachusetts Institute of Technology
A. Test of Self - Gravity Model of Apse Alignment
We have tested the self-gravity model of apse alignment
by comparing its predictions about structure within the
epsilon.ring with an extensive set of observed occultation
profiles covering a wide range of ring longitudes . According
to the self -gravity model of Goldreich and Tremaine (1979) ,
given a radial mass distribution at one ring longitude,
there is a unique distribution of eccentricities across the
ring necessary to maintain locked precession. A specific
prediction for the epsilon ring is that the radial opacity
distribution near periapse and far from periapse would
differ significantly, and in particular that the eccentri-
city gradient across the ring would not be uniform. How-
ever, Voyager occultation observations near periapse (Sigma
Sgr egress) and near quadrature (radio occultation egress)
show very similar shape. The self-gravity model predicts
that the near-periapse profile will have its greatest opa-
city in the inner half of the ring, whereas the observations
clearly show that the near-periapse profile has its greatest
opacity in the outer half of the ring, just as the near-
quadrature profile does .
We conclude that the self-gravity model as presently
constructed is inconsistent with the observations. An addi-
tional strong test of the model will be possible if the sur-
face mass density of the epsilon ring can be estimated. The
model strongly constrains the ring mass by the requirement
that the torque on each ring element be just sufficient to
maintain the locked precession.
B. Lindblad Resonance Survey
During the past year, we determined that the delta ring
has perturbations that are well matched by an m = 2 Lindblad
resonant perturbation (French et al., 1986b). Based on the
information available at the time, we attributed the pertur-
bation to an unseen small satellite inside the orbit of
Miranda. No such satellite was found during the Voyager
encounter, but from analysis of the Voyager trajectory the
mass of Uranus was determined quite accurately (Tyler et
al., 198 6) . They found a mass ratio of the sun to Uranus of
22905.39 + 0.24, whereas we had used Standish and Campbell's
(1984) value of 22951 + 7, which now appears to be in error
by about six times its stated error. The important
30
consequence is that the mean motion of the m = 2 perturba-
tion pattern corresponds to a resonancce radius lying within
a few km of the radius of the delta ring, rather than 41 km
away, as we had concluded from the less accurate Uranus mass
value. It now seems possible that the delta ring exhibits
an internal instability of the sort described by Borderies
et al. (1985) . Whether or not this is the case can best be
settled by determining the ring radii to an accuracy of
about one km, to see if the resonance truly does lie within
the delta ring. This will be done during the coming year.
An addition important test is to see if the Voyager occulta-
tion observations match the perturbation pattern found using
earlier earth-based observations .
C. Shepherd Satellite Ring Perturbations
We are continuing our investigations of variations of
ring width as a function of ring longitude and correlated
these results with radial perturbations. In order to do
this, we will use occultation observations obtained during
May 1985 (French et al., 1985), April 1986, and from Voyager
to map the rings in width and radius . This will be the
major emphasis of our research in the coming year.
D. Ring Orbit Model Enhancement
We have enhanced our kinematical model of the Uranian
ring orbits (French et al., 1986a) to accommodate Voyager
observations as well as ground-based occultation observa-
tions . Most of the past year has been spent on the develop-
ment and careful testing of this major extension to our
existing computer code. The work is now complete, and we
are now able to fit for all ring elements, the direction of
the planetary pole, and J2 and J4, using the complete set of
earth-based and spacecraft observations . We have tested our
results very carefully by writing two sets of independent
code, and cross-checking intermediate results with the Voy-
ager Navigation Team, the radio science group at Stanford,
and with Philip Nicholson at Cornell. We are in the process
of performing a definitive set of orbit fits . Much of our
effort will involve sensitivity studies to determine realis-
tic uncertainties for the derived orbital elements .
References
Borderies, N., P. Goldreich, and S. Tremaine (1985). A
granular flow model for dense planetary rings.
Icarus 63, 406.
French, R. G. et al. (1985). The 4 May and 24 May 1985 oc-
cultations by the Uranian rings. Bull. Amer.
Astron. Soc, 317, 718.
31
French, R.G., J.L. Elliot, and S.E. Levine (1986a). Struc-
ture of the Uranian rings . II . Ring orbits and
widths. Icarus 67, 134.
French, R. G., J. A. Kangas, and J. L. Elliot (1986b). What
perturbs the gamma and delta rings of Uranus? Sci-
ence, 231, 480.
Goldreich, P., and S. Tremaine (1979). Precession of the
epsilon ring of Uranus. Astron. J., 84, 1678-1691.
Standish, E. M. and Campbell, J. K. (1984) . The masses of
the outer planets. Bull. Amer. Astron. Soc, 16,722.
Tyler, G. L. et al . (1986). Voyager 2 radio science obser-
vations of the Uranian system: Atmosphere, rings,
and satellites. Science, 233,79.
32
CHAPTER 2
OUTER PLANETS AND SATELLITES
Shapes of Snail Satellites. P, Thomas, Cornell University
The accurate measurement of limb coordinates on small satellites is
the major technique for evaluating their sizes and shapes in the absence of
good stereoscopic image coverage. The shapes of satellites may be clues to
their internal structure if they are relaxed into ellipsoids (Dermott and
Thomas 5 1986). Small satellites appear not to be relaxed and hence their
shapes may not help in estimating density or moments of inertia. The
distinction between irregular and ellipsoidal satellites can now be made
quantitatively with the availability of many accurate limb profiles of
satellites. The limb profiles are found to subpixel accuracy (Dermott and
Thomas 5 1986) and can be fit by ellipses to approximate an average size and
shape. The residuals from these fit ellipses provide a convenient and
quantitative measure of the topography on the satellites. The standard
deviation of the residuals as a fraction of the satellite's radius gives a
measure of the departure from a smooth form. Departures from gravitational
(and tidal) equipotentials would be greater, but on the average this is a
good measure of the relaxed or non-relaxed nature of the shape of the
object. Figure 1 plots the standard deviation of residuals of limbs for
many small satellites, some of which are averages of several different
views; there are many more data to be included for Phobos and Deimos.
Mimas, a thoroughly studied equilibrium triaxial ellipsoid, is included, as
are preliminary data from the Uranian satellites (Thomas et al,, 1986).
The outstanding feature of this plot is the random distribution of
shapes for objects smaller than Mimas, and the very smooth forms of larger
satellites. The abrupt transition is probably due to a combination of the
effects of the ratio of gravitational binding energy (varies as r^) to the
strength binding (varies as r^) as well as greater relaxation of forms on
larger objects due both to higher gravity as well as slight thermal
gradients within the objects.
The limb topography is also related to specific forms on the objects,
and a systematic survey of these forms on small and larger objects is
underway.
This work is supported by NASA Grant NAGW-lll.
Reference:
Dermott, S. F,, and Thomas, P. (1986). The shape and internal structure of
Mimas. For submission to Icarus.
Thomas, P., J. Veverka, T. V. Johnson, and R. Hamilton Brown (1986).
Voyager observations of 1985U1. For submission to Icarus.
35
10
(0
5 —
- ■ -1 1
1 1
1 1 1 1 II
1
-
—
average limb topography
—
r
s
s
s ■
-
s
—
SB
B
—
—
8
-
_
B a
_
i 1
1 1
a
1 1 1 1 1 1 1*1
1 Bl • 1
-
log radius
Some
Figure. Standard deviations of limb profiles from fit ellipses.
Deviations from equipotential surfaces would be slightly greater,
data points are averages from several pictures as different perspectives
of the same satellite can give very different results. The satellites are
Phobos, Deimos, Calypso, 1980S6, Amalthea, Phoebe, Hyperion, Janus,
Epimetheus, Mimas, and the major Uranian satellites (Uranian satellite
data from Thomas et al . , 1986). Other large satellites would plot even
closer to the X axis than Mimas and Uranian satellites.
36
ACCRETIONAL HEATING OF THE SATELLITES OF SATURN AND URANUS
Steven W. Squyres, Cornel! University, Ithaca, NY 14853; Ray T. Reynolds and Audrey L.
Summers, NASA Ames Research Center; and Felix Shung, Sterling Software, Inc.
Voyager images of the satellites of Saturn and Uranus have shown that these bodies are charac-
terized by remarkable diversity and surprisingly complex geologic histories. Despite their small
sizes, a number of the satellites show unambiguous evidence for resurfacing. Satellites that
have clesirly undergone at least one episode of resurfacing include Enceladus, Dione, Miranda,
Ariel, and Titania [1,2]. Less compelling evidence h^ also been presented for resurfacing on
Mimas and Rhea. In the case of Enceladus and perhaps Ariel, tidal dissipation may have been
largely responsible for heating the satellites. The energy source that apparently led to melt-
ing in the other satellites, however, is less clear. For all the satellites, long-lived radiogenic
heating is entirely inadequate to produce melting. We have therefore turned to investigation
of accretional heating as a possible energy source. Numerous models of accretional heating of
planetary bodies have been developed previously, especially for the terrestrial planets. All such
models suffer froiB a need to select preferred values of critical model parameters in order to
obtain meaningful results. Our goal here has been to develop a detailed model for the heating of
these small satellites, and then to explore the consequences of variations in the free parameters
in the model. Specifically, we attempt to determine for what range of conditions melting will
occur in these satellites. Along with varying a number of model parameters, we also consider
the important effects of inclusion of small amounts of ammonia and (in the case of the uranian
satellites) methane in the system.
We begin with a swarm of bodies from which the satellite will accrete. We allow accretion to
take place gravitationally, calculating the amount of heat deposited, the depth over which it
is deposited, the subsequent transport of heat within the satellite, the rate of growth of the
satellite, and the satellite's evolving internal thermal profile. We consider two reservoirs of
particles from which a satellite may accrete: bodies in orbit about the planet (cis-planetary
debris) and bodies in orbit about the sun (trans-planetary debris) . The initial surface density
ffe of the cis-planetary particle disk is obtained by distributing the present mass of the satellite
over an annular region centered on the planet, with inner and outer radii at the distances at
which the gravitational attraction of the present satellite is balanced by the attractions of its
inner and outer nearest neighbors, respectively. The surface density at of the trans-planetary
disk is determined by similarly distributing a small fraction P of the planet's mass over a sun-
centered annular region. The impacting particles are assumed to have a density equal to that
of the present satellite, and to have a size distribution of the form:
where N{rp) is the number flux of particles of radius fp, and a is a constant.
In calculating the accretion rate and the impact velocity of the impactors from both particle
disks, we follow the formulation of Safronov [3]. We use the Seifronov parameter 6c for the
cis-planetary disk, defined by
e^ = GM,/R,uc^ (2)
and that tke mass flux Me is given by
M.= ^^ii±^..V (4)
"a
where Pg is the satellite's orbital period. Following Kaula [4] , we also define similar expressions
for vt and Mt, the impact velocity and mass flux for debris from the trans-planetary disk.
Having developed expressions for the size distribution, impact velocity, and masB flux of particles
from both sources, we then calculate deposition of thermal energy by the impact process. We
assume that some fraction rj of an impactor's kinetic energy is deposited beneath the satellite's
surface; t] is another free parameter in the model. When an impact takes place, a roughly
hemispheric shock wave centered on the impact point is generated. As it passes through the
material of the satellite, energy is deposited in that material. The specific energy deposited
is larg^t near the impact, and decreases monotonically with distance from the impact point.
We have solved the Rankine-Hugoniot equations for the deposition of specific energy by shock
passage, using published experimental data on shock propagation through H2O ice. The specific
energy e deposited as a function of distance r from the impact point is taken to be
e = eoin/rr (5)
where fj is the impactor radius. Solution of the Rankine-Hugoniot equations for ice show that
an expression of this form is appropriate, and give a value of 7 = 3.1. Energy deposition is
described by eq. (5) out to the point at which the shock pressure drops below the Hugoniot
elastic limit for ice [5], and beyond that point is assumed to be zero. Impacts are Jissumed to
be uniformly distributed over the satellite. Energy deposition is integrated over depth and over
the particle size distribution for both impactor sources, so that the total energy deposition as
a function of depth is obtained.
The satellite is allowed to grow, increasing in radius as it depletes the particle disks from
which it is growing. Simultaneously, we solve the heat conduction equation for the growing
satellite, allowing heat to be transported in response to internal thermal gradients. The thermal
parameters used are those appropriate for H2O ice. Near the surface, heat transport will also
be significantly influenced by the physical mixing that takes place due to impacts, introducing
an additional effective thermal diJQFusivity. The effective diffusivity due to impact mixing at a
depth z is given by
K{z) = ^h,z' (6)
where h^ is the number of times a layer of depth z is turned over by impacts per unit time. We
have developed an expression for K{z) based on the regolith mixing model of Gault et ai. [6].
Growth of the satellite takes place until the source of impactors is exhausted. The free param-
eters in the model are then ^ (the fraction of the planet's mass in the trans-planetary disk),
9c and S^ (Safronov parameters for impactors in both disks), a (the particle size distribution
exponent), and fj (the impactor kinetic energy partitioning coefiicient).
We have performed two sets of calculations. In the first, we have done a large number of
calculations of accretional heating of Saturn's satellite Rhea, allowing for large variations in all
the paramters of interest. A sample result is given in Figure 1, for = 0.001, 6c = 4, 9t = 4,
a = 3.0, rj = 0.2, and 7 = 3.1. Reasonable variations in /3 have a negligible effect on the results.
Even pl£u:ing 10% of the mass of Saturn in the trans-planetary nebula has little effect; the reason
is that the timescale for accretion from the cis-planetary disk is very short in comparison to
that for the trans-planetary disk. Variations in 9c ai^e important. For large values of 9c (i.e.,
38
for low impactor eccentricities), accretion times are extremely short; for example, ~ 200 yr for
6e = 50. Impact velocities are also low, so energy is deposited only very near the surface. In
contrast, low values of 9e lead to longer accretion times (~ 5000 yr for 6c = 2), and deeper
deposition of energy. Variations in a are also important; smaller values of a mean that more
of the mass flux is in large impactors and increases the amount of energy deposited at depth.
As one would expect, variations in f) have a significant effect on the peak internal temperature
achieved.
In the second set of calculations, we have calculated accretional temperature profiles for the
inner major satellites of Saturn (Mimas - Rhea) and for all the major satellites of Uranus. The
maximum temperatures achieved are: Mimas 83 K, Enceladus 93 K, Tethys 145 K, Dione 158
K, Rhea 195 K, Miranda 82 K, Ariel 164 K, Umbriel 160 K, Titania 214 K, Oberon 199 K.
Given the uncertainties in the input parameters, these temperatures may be uncertain by as
much as a few tens of degrees. It is cleair that temperatures will never exceed 273 K, so melting
would not be expected if the satellites were made of pure H2O ice. However, it is probable that
the saturnian satellites contain some NH3, and it is possible that the uranian satellites also
contain some CH4. A peritectic melt will form in the H2O - NHs system at a temperature of
173 K, and it has been suggested that ice containing CH4 could be substantially mobilized at
temperatures as low as 100 K [7]. Given the uncerteiinty in the calculations and the likelihood
of NHs in the saturnian satellites and CH4 in the uranian satellites, these calculations suggest
that, among these satellites, only Mimas and Enceladus cannot have undergone resurfacing as
a result of accretional heating.
References; [1] Plescia, J.B., and Boyce, J.M., Nature 295, 285 (1982); [2] Smith, B.A., et al, Science
2SS, 43 (1986); [3] Safronov, V.S., NASA TT F-677 (1972); [4] Kaula, W.M., J. Geopbys. Res. 84, 999
(1979); [5] Gaffney, E.S., in Ices in the solar system, Reidel (1985); [6] Gault, D.E., et aJ., Proc. Lunar
Sci. Conf. 5tb, 2365 (1974); [7] Stevenson, D.J., and Lunine, J.I., Nature S2S, 46 (1986).
200
0)
3
o
a.
E
F
150
100-
200
300
400 500 600
Radius of Satellite, km
700
800
Figure 1 — Calculated thermal profiles within Rhea during accretion. Curves are plotted at
equal satellite mass intervals. Total accretion time is ~ 2000 yr.
39
INVESTIGATIONS OF THE SURFACES AND INTERIORS OF OUTER PLANET SATELLITES
Gerald Schubert
Department of Earth and Space Sciences
University of California, Los Angeles
Los Angeles, CA 90024
Studies under NSG 7315 during 1985/86 include: 1) Tidal heating and
the structure and evolution of lo and Europa, 2) Crater relaxation on icy-
satellites using realistic non-Newtonian ice rheology, 3) Convection
through phase transitions in the interiors of icy satellites. The
abstracts of published papers on these subjects are reproduced below.
Tidal Heating in an Internal Ocean Model of Europa
M.N. Ross and G. Schubert, Nature, in press, 1986
The tidal response of a three -layer Europa model is calculated. The
model consists of an elastic lithosphere (ice shell) above an invisctd
water layer (internal ocean) and an elastic core (silicate). The tidal
distortion of a decoupled ice lithosphere is a factor of 2 less than
previously thought. For a given dissipation factor Q (equal in the ice
and silicate) tidal heating in the lithosphere is a factor of 4 less than
previous estimates while tidal heating in the core is 30% less. At the
current value of orbital eccentricity tidal heating is only marginally
able to prevent the internal ocean from freezing; Q<25 is required. If Q
in the silicate core equals 100 then Q in the ice lithosphere must be
less than 17 for an internal ocean to be stable against freezing. If
silicate Q=25 then lithosphere Q<37 is required for internal ocean
stability. The persistence of an internal ocean is made somewhat more
probable if a low conductivity surface layer insulates the ice
lithosphere and raises near surface temperatures above the solar
equilibrium value. Cyclical tidal stresses in the lithosphere are
currently about 1/10 of the yield strength of ice, making it unlikely
that these stresses cause the curvilinear features on Europa.
Finite Element Models of Non-Newtonian Crater Relaxation
P. Thomas and G. Schubert, Proc. Lunar Sex. Conf. 17th,
J. Geophys. Res., in press, 1986.
Models of viscous crater relaxation proposed to date for the icy
satellites of the outer solar system assume the behavior of a Newtonian
mediiom, with viscosity rf independent of effective stress r. While this
is reasonable if both effective stresses and temperatures are very low,
laboratory data on ice indicate that, at effective shear stresses (r ~
0.1 MPa) and temperatures (~ 173K) likely to occur in regions underlying
craters, non-Newtonian behavior is probable. Such material would have a
viscosity-effective stress relation given by f?(r)oc t^'", where n » 4. It
is likely, due to the low activation energy of ice at these temperatures.
40
that stress differences may affect the profile of relaxing craters to a
greater extent than temperature differences . To investigate the effects
of a non-Newtonian rheology on the profiles of relaxing craters, two-
dimensional finite element simulations have been performed. Initially,
effective stresses are highest in the region below the crater center,
producing a zone of relatively low viscosity. Relaxation flow forces this
region into a convex bulge as the crater relaxes. At later stages, the
low viscosity region moves to beneath the crater rim, leading to enhanced
relaxation of crater rims, compared to a Newtonian rheology. The
reduction of effective stress as the crater relaxes increases viscosity
in the region beneath the crater. As a result, relaxation rates are
quite rapid compared to those in later stages. This is in marked
contrast to the exponential behavior of depth with time associated with
Newtonian relaxation. The short relaxation times observed for the
craters modelled indicate that a silicate component may be required in
the crusts of icy satellites to account for the observed crater
populations .
Tidally Forced Viscous Heating in a Partially Molten lo
M.N. Ross and G. Schubert, Icarus, 64, 391-400, 1985.
We investigate tidal dissipative heating in two different models of
lo. The partially molten asthenosphere model consists of a rigid core
and a thin (less than 400km thick) partially molten "decoupling" layer
(asthenosphere) surrounded by an elastic lithosphere. In the partially
molten interior model the interior beneath the lithosphere is partially
molten throughout. The partially molten region in each model is assumed
to possess negligible shear strength and to be characterized by a
Newtonian viscosity. Tidal deformation and dissipation in the core of
the thin asthenosphere model are assxamed negligible. Fluid in the
viscous layers is forced to circulate by the tidal distortion of the
outer shell, modeled here as a sinusoidal variation with time of the
distortion amplitude. As a result, heat is generated in the fluid by
viscous dissipation. There are two heating mechanisms in our models:
"elastic" dissipation in the lithosphere ex 1/Q and viscous dissipation in
the partially molten region. Numerical calculations are carried out for
a 90 -km- thick lithosphere with Q=100. This thickness maximizes
dissipation in a decoupled lithosphere; other reasonable values of
lithosphere thickness do not alter our calculations. Under the
constraint that total dissipation equals the observed radiated heat loss
we derive the viscosity of the partially molten region in each model. We
a posteriori evaluate the assumption that the lithosphere is decoupled
from the interior by calculating the distortion of an elastic shell due
to the viscous stresses on the lower surface of the outer shell. If the
interior viscosity is such that the total dissipation is equal to the
observed heat flux from lo, viscous stresses produce negligible
distortion of a 90 -km- thick shell. This validates the assijmption of a
decoupled shell. The derived viscosity for both models is characteristic
of a partially molten rock. In the thin asthenosphere model the derived
41
viscosity is so low that a very high degree of partial melt is necessary,
about 40% crystal fraction in a 400 -km- thick asthenosphere and about 0%
in a 1 -km- thick asthenosphere. In the partially molten interior model
the derived viscosity corresponds to a magma with about 60% crystals.
Consideration of convectlve efficiencies demonstrates the plausibility of
a stable thermal steady state for both models. A significant portion
(75% for Q = 100) of lo's tidal heating can be the result of viscous
dissipation in a partially molten region that decouples the outer shell
from the interior. The partially molten layer can be considered a
"global magma ocean."
Phase Translstions and Convection in Icy Satellites
D. Bercovlcl, G. Schubert, and R.T. Reynolds,
Geophjs. Res. Lett., 13, 448-451, 1986.
The effects of solid- solid phase changes on subsolidus convection in
the large icy moons of the outer solar system are considered. Phase
transitions affect convection via processes that distort the phase change
boundary and/or influence buoyancy through thermal expansion. Linear
stability analyses are carried out for ice layers with a phase change at
the midplane. Two exothermic phase transitions (ice I - ice II, ice VI -
ice VIII) and two endothermic transitions (ice I - ice III, ice II - ice
V) are considered. For the exothermic cases, the phase change can either
Impede or enhance whole-layer convection. For the endothermic cases, the
phase change always inhibits whole -layer convection overturn and tends to
enforce two -layer convection. These results place some constraints on
possible models of icy satellite evolution and substructure.
Crater Relaxation as a Probe of Europa's Interior
P.J. Thomas and G. Schubert, Proc. Lunar and Planet. Sci. Conf. 16th
Planet. Sci. Conf. 16th, J. Geophys. Res., 91, D453-D459, 1986
Viscous relaxation rates of craters are examined to discriminate
among models of Europa's interior. It is shown that the presence of
an insulating surface regolith is required for ice lithospheres that
do not contain a liquid HgO layer. Without the presence of such an
Insulating regolith, relaxation times for craters - 100 km in
diameter are excessively long ( > 1 b.y.) and thus inconsistent with
the absence of such features in Voyager images. Finite element
calculations indicate that a thick insulating surface frost layer
(raising surface temperatures > 40 K above the solar radiation ambient
temperature of 92 K) allows a 25 -km- thick surface viscous layer to
relax > 100 -km- diameter craters on the short timescales (- 100 m.y.)
required. If the icy lithosphere is 100 km thick (and without a
liquid HgO layer, but containing a subsolidus convectlng region),
surface temperatures need to be raised 20 K by Insulation to allow
relaxation of craters in this size range to occur within the required
time.
42
FAILURE STRENGTH OF ICY LITHOSPHERES
M. P. Golombek and W. B. Banerdt (Jet Propulsion Laboratory, Caltech,
Pasadena, CA 91109)
It has been widely assumed in the literature that the stresses required
for brittle failure of an icy lithosphere under tension are limited by the
tensile strength of intact ice. This assumption has led to the widespread
use of 2 MPa (20 bars) for the extensional strength of predominantly ice
lithospheres (e.g. Ganymede's), based on the unconfined tensile strength of
ice near its melting temperature measured by Hawkes and Mellor (1). However,
our understanding of the maximum stress levels in the earth's lithosphere
are based on the frictional resistance to sliding on pre-existing fractures
and not the strength of- intact rock in laboratory experiments, which is
always greater. Similar friction relations for ice predict maximum stresses
that are substantially greater than the apparent tensile strength of intact
ice near its melting temperature. At first inspection, then, it appears that
ice rich lithospheres would fail under tensional stress by tensile fracture
of intact ice rather than sliding on pre-existing fractures. In this ab-
stract we will introduce lithospheric strengths derived from friction on
pre-existing fractures and ductile flow laws, derive these relations for icy
lithospheres, show that the the tensile strength of intact ice under appli-
cable conditions is actually an order of magnitude stronger than widely
assumed, and demonstrate that this strength is everywhere greater than that
required to initiate frictional sliding on pre-existing fractures and
faults .
The maximum stress levels found in the earth's crust are accurately
predicted by Byerlee's law (2,3). This relation is based on laboratory
measurements of the frictional resistance to sliding on pre-existing frac-
tures, which occurs at stresses less than those required to break intact
rock. Byerlee's law is of the form t=uS^+b, where t is the shear stress, S^^
is the effective normal stress (normal stress minus pore pressure), u is the
coefficient of friction, and b is a constant. In terms of S-j^ and S3, the
maximum and minimum principal effective stresses (stress minus pore pres-
sure), Byerlee's law can be written S-|^=KS3+B, where K=[(u^+l)-'-/^+u] ^, B is
a function of b and u, and sliding is assumed to occur on the most advanta-
geous slip plane. Laboratory friction measurements show that u and b are
virtually independent of stress (except for a slight change for rocks at 135
MPa), rock type, displacement, surface conditions, and temperature. For this
application we will use the friction law for low stress determined by
Byerlee (2) t=0.85S^, which has been found to hold for a wide variety of
geologic materials, because the appropriate low normal stress friction
measurements on ice have not been made. At higher stress we use the friction
data measured for ice, u=0.2 and b=10 MPa (4). The higher stress friction
data for ice do constrain the applicable low stress friction law to be very
close to that for rock, because the lowest normal stress measurements for
ice are for S^=17 MPa (5), which constrains the low stress friction law,
that also must pass through the origin, to have a slope only marginally
different (u>0.79) from that determined for rocks (u=0.85). In order for
fluid pore pressure to decrease the friction on pre-existing faults and
fractures, thereby changing the slope of the brittle yield stress versus
depth curve, liquid water must fill the connected pore space. This is not
likely to have occurred on the icy satellites given the extremely low
surface temperatures (less than 100°K) and reasonable thermal gradients.
Because the vertical stress is generally quite close to the lithostatic load
(density times gravity times depth) in areas of low relief, this relation
43
predicts a linear increase in yield stress with depth (Fig. 1).
With increasing temperature, rock and ice deformation occurs by ductile
flow. Flow laws for rocks, minerals, and ice have been experimentally deter-
mined for stresses up to 1-2 GPa and strain rates down to 10' /sec. These
results can be extrapolated to geologic strain rates via creep equations
which generally are of the form de/dt=A(S2^-S3)'^exp(-Q/RT), where de/dt is
the strain rate, R is the gas constant, T is absolute temperature, and A, Q
(the activation energy), and n are experimentally determined constants. As a
result the ductile strength is negligible at depths where T is high and
increases exponentially with decreasing depth. We have used the experimen-
tally determined flow parameters (6,5) for pure ice I^^, A=1.2xlO" /sec- Pa ,
Q=4.5xl0 J/mole and n=4.0 extrapolated to geologic strain rates of 10" /sec
(about 3%/m.y.). A surface temperature of 100°K and a thermal gradient of
1.6°/km applicable to the early high temperatures likely in some of the
larger icy satellites (e.g. Ganymede, 7) yields the strength envelope illu-
strated in Fig. 1. It has been shown at high temperatures (i.e. polar
conditions on the earth) that the inclusion of a small amount of silicates
will greatly increase the creep strength of ice 1-^^ (8). The amount of
hardening from the inclusion of small amounts of silicates in the litho-
spheres of the icy satellites is not known, but for lack of a better con-
straint we will assume that the hardening resulting from the addition of
less than a few percent of silicates in icy satellites lithospheres can be
bracketed by an order of magnitude increase in creep strength (e.g. 9).
The failure criterion for a given depth in the lithosphere is deter-
mined by the weaker of the frictional or ductile strength at that depth. The
yield stress increases with depth according to Byerlee's law until it ex-
ceeds that calculated using the appropriate flow law, after which it de-
creases exponentially with depth. 'The intersection of the brittle and duc-
tile yield stress curves defines the brittle-ductile transition depth and
also the peak stress needed to cause failure of the entire lithosphere. As a
result the peak stress needed for lithospheric failure is dependent on the
thermal gradient, which causes the ductile flow law to intersect the fric-
tion curve at different depths (shallow for high thermal gradients, deep for
low thermal gradients). These peak stresses are probably somewhat greater
than those actually required to cause failure of the lithosphere because
semibrittle and low temperature ductile processes tend to round off the
intersection points between the brittle and ductile curves (10).
As can be seen in Fig. 1, the peak stress needed to cause tensile
failure of the lithosphere using the above parameters is on the order of 10
MPa (100 bars), although the average stress in the lithosphere is only about
half that value. This strength is applicable for large icy satellites (e.g.
Ganymede). We have also determined strengths for the smaller icy satellites
of Saturn using the appropriate gravities, densities, surface temperatures,
and calculated thermal profiles (11) and found all strengths are substan-
tially larger than this value. As a result this value is probably a minimum
for the lithospheres of the icy satellites.
Note that this peak stress (10 MPa) is significantly higher than the
tensile strength of about 2 MPa used in virtually all previous studies,
which is based on the the unconfined tensile fracture strength of ice near
its melting temperature. Because failure should occur by the mechanism
requiring the lowest stress, it would appear on first inspection that ten-
sile fracture of intact ice is the relevant mode of fracture for icy satel-
lites. However it can be shown using a Griffith failure criterion and taking
the vertical lithostatic load into account that the stress difference for
failure due to the tensile fracture of intact material at depth z is given
44
by S3^-S3=SQ+pgz for S;l-S3<^So and by S]^-S3=4[ (SQpgz+S^j2)l/2.s^] for
S-|^-S3>4Sq, where p is density, g is gravitational acceleration, and S is
the unconfined tensile strength (e.g. 12). The first of these relations des-
cribes the opening of tension cracks, which occurs when the confining pres-
sure is relatively low. The second relation describes shear failure in
tension (or compression), which is the mode of failure when the confining
pressure is too great for open tension cracks to form. The depth of transi-
tion between the two modes of tensile failure is given by z=3SQ/pg (about 5
km for Sq=2.5 MPa on a satellite with a gravitational acceleration similar
to that of Ganymede). The curve for failure due to fracture of intact ice is
plotted in Fig. 1, where it can be seen that frictional failure is preferred
at all depths for the assumed material parameters of ice. This conclusion is
borne out by the experimental results, in which ice samples failed due to
frictional sliding along pre-existing saw cuts rather than initiating new
fractures (4). Note also that if the tensile fracture strength of intact ice
increases with decreasing temperature, as is the case for the compressional
fracture strength (13), the curve for fracture of intact ice under tension
will move to the left and frictional failure will be even more favorable.
Similar results are also obtained for other choices of failure criteria.
Thus the failure strength of an icy lithosphere is significantly greater
than has been previously assumed.
In conclusion, because the tensile strength of intact ice increases
markedly with confining pressure, it actually exceeds the frictional
strength at all depths. Thus, icy lithospheres will fail by frictional slip
along pre-existing fractures at yield stresses greater than previously
assumed rather than opening tensile cracks in intact ice.
References (1) Hawkes & Mellor (1972) J Glaciol 11, 103. (2) Byerlee (1978)
Pageoph 116, 615. (3) Brace & Kohlstedt (1980) JGR 85, 6248. (4) Beeman et
al. (1984) EOS 65, 1077. (5) W. Durham, written com. (6) Durham et al.
(1983) P14LPSC JGR 88, B377. (7) Golombek & Banerdt (1986) Icarus Nov. 86.
(8) Baker & Gerberich (1979) J Glaciol 24, 179. (9) Friedson & Stevenson
(1983) Icarus 56, 1. (10) Kirby (1980) JGR 85, 6353. (11) Ellsworth &
Schubert (1983) Icarus 54, 490. (12) Jaeger & Cook (1976) Fund. Rock Mech.
(13) Parameswaran & Jones (1975) J Glaciol 14, 305.
Fig. 1. Brittle and duc-
tile yield stress versus
depth curve for compres-
sion (to the right) and
extension (to the left),
with a peak stress of 10
MPa (see text for discus-
sion). The dashed lines
show the failure strength
of intact ice as a func-
tion of depth assuming a
Griffith failure criteri-
on. Because stresses re-
quired to break intact
rock are greater than
those required to initiate
sliding on pre-existing
fractures, lithospheric
failure occurs by fric-
tional sliding on pre-
existing fractures.
E
a.
O
-10
"'C
~^-
-20
\l
If
-30
^
1
-40
1
-20
— 10 O 10 20
Stress Difference (MPa)
30
45
STUDIES OF OUTER PLANET SATELLITES, MERCURY, AND URANUS
William B. McKinnon and Paul M. Schenk, Department of Earth and Planetary Sciences and
McDonnell Center for the Space Sciences Washington University, Saint Louis, MO 63130.
Ring Geometry on Ganymede and Callisto
Arguments have been made, based on geometry, for both an impact and an internal origin for the
ancient, partially preserved furrow system of Ganymede. Zuber and Parmentier (1984) concluded
that furrows were not concentric, but could be impact related if multiringed structures on icy satellites
are initially noncircular. We examine the geometry of the Valhalla ring structure on Callisto in order
to assess the circularity of an unmodified ring system. Despite prominent local meandering, the only
gross deviations from concentricity in the Valhalla system are found in the outer north-east quadrant
of the system. Here, a number of ring segments intersect small circles about the center at angles up
to 30° . The Ganymede furrow system was remapped to make use of improvements in coordinate
control. The least-squares center of curvature (determined using natural weighting) for all furrows in
Marius and Galileo Regio is -20.7°, 179.2°. Furrows in Marius and Galileo Regio are reasonably
concentric, and are much more circular than previously estimated and probably once covered at least
an entire hemisphere of Ganymede. Thus we find furrow geometry is consistent with an impact
origin (Schenk and McKinnon, 1986, 1987). Deviations of some furrows from concentricity about
the center of curvature, on the scale of those found at Valhalla, do exist. As in the case of Valhalla
these variations are principally confined to outer regions of the structure, and are interpreted as
inherent properties of multiringed structures on icy satellites. The cause(s) of this may be in the ring
formation mechanism itself, but are more likely due to variations in preexisting lithospheric
mechanical properties. The perceived present nonalignment of the assumed originally concentric
furrows has been used to argue for large-scale lateral motion of dark terrain blocks in Ganymede's
crust, presumably in association with bright terrain formation. The overall alignment of furrows as
well as the inherent scatter in centers of curvature for subregions of Galileo and Marius Regio do not
support this hypothesis.
Primitive Material on Ganymede and Callisto
The surfaces of the two outermost Galilean satellites may record the influx of carbonaceous and
"ultracarbonaceous" material in the Jupiter region (McKinnon and Schenk, 1986). Major albedo
units on both satellites are spectrally red as determined by Voyager (excepting the brightest craters on
Ganymede), and increasing redness is correlated with decreasing albedo and increasing crater density
(age). Present optical, infrared, and microwave data indicate very ice-rich optical surfaces and
regoliths, globally averaged, for both bodies. Together, these argue for the presence of an exogenic,
dark, reddish, contaminant. Regionally, the darkest and reddest units on Ganymede are dark-ray
craters, although they are not as dark as lapetus dark material or even as dark as Callisto dark terrain.
In some cases they are no more red than nearby dark terrain, and one large dark-ray crater, Kittu, is
actually spectrally neutral (in the visible). If the rays are areal mixtures of ice-free and ice-rich
materials, then the non-ice portion may in most cases be analogous to D-type "ultracarbonaceous"
material, and in the case of Kittu, "ordinary" C-type material. The best model for formation of dark
rays involves the creation of a lag deposit in the ejecta of projectiles of rare low velocity or
composition or both (Conca, 1981). We infer the compositional class to be (mostly) either D-class
asteroids or comets (possibly silicate-enriched). We cannot rule out the impact of C asteroids coupled
with spectral modification by the Jovian magnetosphere, but Kittu makes this hypothesis less likely.
Terrain contamination by the integrated effects of dispersed D asteroid ejecta and infalling D-type
meteoritic dust is also plausible, if accumulation preceded crater retention in the heavily cratered
terrains. D-type objects are also the most likely ones available in the Jovian region; C types may
have only dominated Jupiter-crossing objects during dispersal of the main asteroid belt by Jupiter,
46
apparently near the time the protojovian nebula Itself dispersed if the flat spectra of the "captured"
outer Jovian satellites are to be explained. "Himalia-dust" alone cannot account for the contamination
of the surfaces of Ganymede and Callisto without spectral alteration. Work on this topic Is
continuing.
Tectonics of Colons Basin, Mercury
Caloris is in many ways a unique basin compared with its lunar counterparts (McKinnon, 1986a).
As a multiringed basin . Although youthful and vast, most of its multiringed elements are obscured.
Ring segments imniediately outside the Caloris Montes lie at diameter ratios smaller than 1 .4. If the
ridge systems and concentric arches of the basin floor indicate buried rings, their spacings are
consistent with the inter-ring distances outside the basin rim and indicate an upper limit to the
thickness of the impact-defined lithosphere of -120 km. As a volcanic center . The smooth plains
fill, presumably volcanic, is extraordinarily deep if it is required to completely bury ring topography;
alternatively, the mercurian lithosphere may not be as buoyant as that of the moon. Later lunar basin
volcanism usually occurs near the basin periphery; at Caloris this apparently corresponds to the
emplacement of the smooth plains in an annulus surrounding the basin. That this annulus may be
nearly complete is supported by radar and earth-based visual observations. All these volcanic
manifestations provide a model for intercrater plains formation. The only lunar analogue to this
volcanic style might be the Oceanus Procellarum basalts, if they are viewed as a partial annulus
surrounding the Imbrium basin. As a mascon . Topographic depression and ridge formation imply
that the basin floor subsided under a load, but the predominantly concentric orientation of the ridges
is not consistent with a central mascon (radial ridges are predicted) unless basement ring structure
controls tectonic orientation or the scale of the load is larger than Caloris itself. The latter may be
consistent with a broad withdrawl of magma to form the smooth plains, as advocated by Dzurisin
(1978). The final tectonic episode appears to be uplift and extension of the basin floor, and thus any
present mascon is likely to be associated with the exterior smooth plains. The effects of such a ring
load were investigated (McKinnon, 1979) using the thick-plate theory of Melosh, in which shell
curvature is parameterized. The main characteristics of the predicted tectonic pattern are normal
faulting within the basin and thrust faulting beneath the ring load, both in agreement with
observation. The dominant concentric trend of the basin norma! faults is consistent with the ring load
hypothesis provided the elastic lithosphere of Mercury was ^ 125 km thick at the time of faulting.
Although this solution is approximate due to the great scale of the structure compared to Mercury's
radius, it suggests that central concentric normal faults are characteristic of annular loading in the
long-wavelength limit. (The dominant long-wavelength component of the annular load is second
order, and thus equivalent to the tidal despinning problem. Concentric normal faults within the basin
are predicted in this case.) Simple updoming within the basin would produce normal faults of
predominantly radial orientation. A lower limit to the elastic lithosphere thickness (~75 km) is set by
requiring sufficient deviatoric stress to initiate faulting for reasonable load magnitudes (consistent with
photogeological observation, radar subsidence profiles, and limits to the contribution to Jj).
Microwave Interferometry of the Deep Atmosphere of Uranus
A less-appreciated feature of the Uranian microwave spectrum is its apparent flatness at
wavelengths longward of ~6 cm. From the suite of measurements made by various workers during
1978-1979 (the Uranus spectrum can be time-variable), averaged brightness temperatures at 6, 13,
and ~20 cm can be estimated as 251 ±5, 255±18, and 265±45 K. The last figure Includes a strenu-
ous attempt to measure the 21 -cm brightness temperature with the Owens Valley three-element
interferometer. Eighty hours of integration over five days yielded a value of 240±63 K (McKinnon et
al., 1981). More recent higher-precision VLA measurements by Jaffe et al. (1984) and de Pater
(Gulkis and de Pater, 1984) give brightness temperatures ranging between ~225-235 K at 6 cm and
~240 at 21 cm, suggesting that the averages above -were biased upward or that the decimeter flux has
fallen in recent years. In either case there is no evidence that the spectrum is steeply rising with
47
wavelength, and it may well be flat. We originally considered three explanations: a subadiabatic
region, a high surface, and a rapid increase in microwave absorption below some level. The first two
were rejected on physical grounds. I examine the last by calculating theoretical brightness
temperatures by standard methods, incorporating the Voyager 2 radio occultation profile as the upper
boundary condition (McKinnon, 1986b). One class of model atmosphere has only molecular
hydrogen and water vapor as opacity sources. (This may actually occur if a super-massive water
cloud [>several 100 x solar adundance] traps essentially all the available ammonia deep in the cloud.)
As long as the water abundance is greater than a few times solar, then the calculated spectrum is
rather flat, ranging between ~270 and ~290 K as the wavelength varies from 6 to 21 cm. Although
this spectrum is two warm to fit the decimeter data, emission ffom a water-cloud deck may account
for the 270 K warm regions seen in 6-cm VIA maps. The decimeter brightness spectrum requires
that the water abundance of the deep atmosphere not be so great that a sensible amount of ammonia
vapor is not distributed between the base of the water-ice cloud and the top of the probable NH^ SH
cloud. Including ammonia in the brightness temperature modeling, using abundance profiles
calculated by S.K. Atreya (in the manner of Atreya and Romani [1985]) and Fegley and Prinn
(1985), shows that disk-averaged brightness temperatures in the 6-21 cm range are well fit by a deep
atmosphere with abundances of water and ammonia of ~20 x solar. In this sense, ammonia has been
detected on Uranus. Any global abundance determination is suspect, however, because the disk-
resolved VLA maps show strong latitudinal variation in ammonia abundance. Perhaps the polar
regions, which are ammonia-free at the wavelengths discussed here, are more representative of the
deep atmosphere, thus implying a much greater enhancement in water and ammonia (> several 100 x
solar) at deep levels. The detectable ammonia in the Uranian atmosphere at equatorial and mid-
latitudes may be due to convective lofting.
Ackowledgement . Research on outer planet satellites is supported by NASA grant NAGW-432.
Research on Mercury and Uranus is a continuation of long-standing work, and is independently
supported.
References
Atreya S.K. and P.N. Romani (1985) In Recent Advances in Planetary Meteorology , 17-68; ConcaJ.
(1981) Proc. Lunar Planet. Sci. Conf . I2B, 1599-1073; Dzxdrisin D. (1978) J. Geophys. Res . 83, 4883-
4906; Fegley B. and R.G. Prinn (1985) Nature 318, 48-50; Gulkis S. and I. de Pater (1984) In Uranus
and Neptune, NASA Conf. Pub. 2330 , 225-262; Jaffe W.J., G.L. Berge, T. Owen and J. Caldwell (1984)
Science 225, 619-621; McKinnon W.B. (1979) Eos Trans. AGU 60, 871; McKinnon W.B. (1986a)
Mercury Conference Abstracts , 18; McKinnon W.B. (1986b) Bull. Am. Astron. Soc . 18, 765; McKinnon
W.B. and P.M. Schenk (1986) Eos Trans. AGU 67, 1073; McKinnon W.B., G.L. Berge and D.O.
Muhleman (1981) Eos Trans. AGU 62, 941; Schenk P.M. and W.B. McKinnon (1986) Bull. Am. Astron.
Soc. 18, 759-760; Schenk P.M. and W.B. McKinnon (1987) Icarus , submitted; Zuber M.T. and E.M.
Parmentier (1984) Icarus 60, 200-210.
48
ms ON 10: A THEEMOD-YHAMIC PEISPECTWE
Aaron P. Zent and Fraser P. Fanale, Planetary Geosciences Division, Hawaii Institute
of Geophysics University of Hawaii, Honolulu, Hawaii
The presence of condensed SOg on lo mandates a finite abundance of SOg vapor
which must be present, regardless of plume activity. Currently, even the order of
m.agnitude of the ambient SOg pressure is unknown. However, a number of models
indicate that the pressure may be near saturation much of the time. Among the
models that suggest atmospheric pressure does not approach saturation is that of
Matson and Nash. [1983], who base their model in part on the argument that lo's
surface must have a very high porosity, and in part on the argument that the parti-
culate material in lo's volcanic plumes has a very fine grain size. This implies that a
significant amount of cold, particulate, surface area is likely to be in diffusive con-
tact with an SO2 atmosphere that probably at least approaches saturation locally.
These conditions indicate that adsorption of SOjj may be important SOg adsor-
bate. We hav® measured the adsorption of SOg on particulate sulphur, and exam-
ined the equilibrium between adsorbed SOg, SOg vapor, and SOg ice based upon our
measurements and simple thermodynamic considerations.
Measurements of SOg adsorption on particulate sulphur at conditions near
those of lo were made on a Numinco Model MIC 103 OOR Surface-Area Pore-Volume
Analyzer. Adsorption data were gathered at 178K, 193K and 225K. These are higher
than surface temperatures on lo, which peak around 130K. The SOg vapor pressure
(1 to lO^Pa) also exceeds lo surface pressures. The adsorptive behavior of the sys-
tem must be extrapolated to lo conditions.
The most rigorous extrapolation method of which we are aware is one
presented by Anderson et al. [1987]. They showed that it is possible to compute
adsorption isotherms for temperatures below those covered in adsorption measure-
ments via the equation
In
(PVP)g
(PVP)i
(AHva-AH^)(T3-Ti)_
RTiTg
Where P^ and P* are the vapor pressures over the adsorbed phase and the solid
phase, respectively. The subscript 1 refers to state variables at condition 1, where
data are available, and 2 refers to state variables at condition 2 for which we would
like to calculate the relative vapor pressure. AHya is the change in partial molar
enthalpy in going from the vapor to the adsorbed phase, and AHyj is the same quan-
tity for the vapor to ice phase change.
In order to use Eq.(l), we must know AHvj(T) and, from the data, we must find
AHva(T). AHva is found by solving the Clausius-Clapejron equation at points of con-
stant adsorptive coverage. Since the data are not at constant adsorptive coverage
we must fit a curve to the data and read equilibrium pressures at constant adsorp-
tive coverage from that.
We are then able to extrapolate our adsorption isotherms to lower tempera-
tures via Eq.(l). In all cases, our isotherms agree with the data to better than a
factor of 2.5; the average precision of the data is ±50%.
At a fixed temperature, the equilibrium vapor pressure over adsorbed SOg may
be thought of as an extensive property of the system; it increases as the mass of
SOg in the system increases. The equilibrium vapor pressure over ice however is
intensive; it does not matter how much SOg is in the solid phase, the vapor pressure
depends only on temperature. That means that at a fixed temperature, there is a
maximum adsorptive coverage, which occurs at the point at which the equilibrium
49
vapor pressure over the adsorbed phase is equal to the vapor pressure over ice.
Thereeiter, any additional SOg present in the system must exist as ice.
We should now be able to calculate the m.aximum adsorptive coverage in terms
of volume of SOg per unit m.ass of sulphur, although we consider this an intermedi-
ate data product. ¥e would really like to know the maximum coverage of SOg in
terms of number density per unit surface area of sulphur (i.e. the monolayer cover-
age). We need to represent our results in this way because the specific surface area
of S on lo is unconstrained, but the upper lunit on the number density uf the cover-
age, which is more closely related to the chemical potential, must be the same in
the laboratory as on lo.
Fortunately, adsorption measurements provide concurrent measurements of
specific surface area through application of BET theory [Brunauer et al., 1938]. We
calculate BET surface areas, assuming the size of the adsorbed SOg molecule is ~
30A^. We use the mean specific surface area of ~ 5 x 10"^ m^ g~' indicated in otir
analysis to calculate maximum coverage for sulphur at lo conditions.
This relationship is shown in Figure 1, along with the data. Each of the isoth-
erms is truncated gainst the vapor pressure curve for SOg ice (solid line). A very
simple expression can be written to describe the maximumi amount of adsorbed SOg
per square meter of sulphur, at any temperature.
Max Pa = exp(aT-b) (2)
where a = 0.1673 K~\ and b = 33.89, and Pe is the density of adsorbed SOg in terms
of g SOg m~^S. If the maximum surface temperature on lo is around 130 K, the max-
imum abundance of adsorbed SOg on lo is around 5 x 10~® g SOg m~^ S. At disk aver-
age temperatures of 90 K, the maximum adsorbed SOg capacity cannot be more
than 8.6 X 10"^ g SOg m.~^ S. If more SOg is present in the system at these tempera-
tures, it will exist as ice.
To understand why our results constitute a new perspective, we must examine
the previous estimates of SOg adsorption on lo.
Unpublished adsorption measurements of SOg on S obtained by Fanale and
Laue, at warmer, higher pressure, conditions suggested that up to a monolayer of
SOg could be adsorbed onto S at low relative pressures. It was more or less
accepted that at lower pressures and temperatures characteristic of lo, the adsorp-
tive coverage should be about the same, around a monolayer [e.g. Fanale et al.,
1982]. Let us examine how many monolayers our current analysis indicates. The
upper limit of adsorptive coverage is around 5 x 10~^ g SOg m~^sulphur at 130K (Fig.
1). This mass represents around 4.7 x 10^^ molecules per square meter. If we
assume that each adsorbed molecule occupies 30A^, the maximum number of
adsorbed SOg molecules occupies no m.ore than 1.4 x 10~^ m^. Thus ~ 1.4 x 10~^ is
the maximum number of adsorbed monolayers, well over an order of magnitude less
than previously believed. Nash [1983] calculated that 0.5 monolayers of adsorbed
SOg would be required to create the 30% 4— /im absorption. We argue that the max-
imum adsorbed SOg coverage must be significantly less than 0.5 monolayers, mak-
ing spectroscopic detection of adsorbed SOg highly problematic.
The question of how important the adsorbed phase on lo is, now becomes
largely a question of what the adsorbent is.
We selected sulphur as our adsorbent because it is clear that sulphur allo-
tropes constitute a significant fraction of lo's surface. To the extent that elemental
sulphur dominates the Ionian surface, we have excellent confidence in our results.
However, the Na and K in the magnetosphere require the presence of other phases
on the surface.
50
The i mpor tance of the adsorbed phase in suppling the magnetosphere, and in
diurnal exchange of SOg with the atmosphere, depends on the abundance and
adsorptive capacity of other surface materials, as well as on ihe specific surface
area of the sulphur on the lo surface. Nash [1986], in experiments on vacuum subli-
m.ation of solid sulphur, identified a very fluffy form which results from preferential
sublimation of ring sulphur (Sa) over polymeric sulphur. Such a filamentary residue
might be expected to have a very high surface area, and hence render the surface
adsorbed phase a significant sink by mass, although monolayer coverage would
remain very small.
REFERENCES
1. Anderson, et al., Science. 155. 319-322. 1967.
2. Brunauer, S., et al., /, Am. Chem. Soc. 60, 309-319, 1938.
3. Fanale, F. P.. et al.. In: Satellites of Jupiter (Morrison, D., ed.), 756 - 781.
4. Matson, D. L.. and Nash. D. B... /. Geaphys. Res.. 88, 4771-4783, 1983.
5. Nash, D. B., Icarus, 54, 511-523, 1983.
6. Nash, D. B.. In: Lunar and Planetary Science. XVII, pp. 603-604, 1986.
1982.
'E
CM
o
C/)
CO
o
to
Q
<
1«!0'^-
8siO"*-
dKlO"*-
4x10
2x10'
1«10
1«10'
Pressure (Po)
1x10'
Is 102
la 103
Pig. 1 - The pressaire temperature field tangential to that of lo. The isotherms are found from Eq. (1)
(daahed lines) and each is truncated at the ?apor pressure over SOg ice (solid line). Coverage
corresponding to one monolayer is indicated by the arrows (®). On lo, (T< 130K), adsorption is limited to
& fraction of a monolayer hy equilibration mHi ice.
51
INVESTIGATIONS OF PLANETARY RING PHENOMENA
Joseph A. Burnsj Cornell Universitlly
Faint planetary rings — their dynamical behavior and
physical properties — have been the main focus of research
efforts during the past year. A reanalysis of the structure
and physical properties of the Jovian ring system was
completed and will be published shortly (2). Jupiter's ring
is composed of i) a flattened main ring (t<^300 km) that
terminates abruptly at its outer edge, and that has features
associated with the two known embedded moonlets, ii) a
toroidal halo that rises quickly at the main ring's inner
edge to a half thickness of '^10'* km and fades into the
background at 't'l . 4 Rj , and iii) the exterior "gossamer" ring
(see last year's progress report) that is only 1/20 as bright
as the main ring and that is enhanced at synchronous orbit.
The ring is reddish like nearby Amalthea and its particles
have a power law size distribution of slope -2.5(+0.5) in the
range -^0 . 1-lOO^m.
We have examined the motion of weakly-charged dust
through Jupiter's gravitational and magnetic fields
(1). Resonances occur where the frequency of the perturbing
Lorentz force matches a grain's natural orbital frequency;
near such resonances, large radial and out of plane motions
appear. Ongoing research shows that during passage through
such resonances large vertical amplitudes develop and
persist; the boundaries of the jovian halo are near such
locations. Current studies include the character of
stochastic variations in a particle's electric charge and the
orbital consequences of this, an analytical investigation of
resonance passage, the details of the dust-plasma
interaction, and application to other ring systems.
Several topics concerning features of Saturn's rings
have been addressed. Cuzzi and Burns (4) have shown that
depletions in MeV charged particles measured by Pioneer 11
near the F ring are only partly caused by the ring and thus
that unobserved localized clouds of debris must be
present. We hypothesize that these faint clouds are ejecta
thrown off during collisions amongst moonlets (radius 0.1-10
km) which populate the entire annulus between the orbits of
the shepherds. We have developed a self-consistent scenario
in which mutual collisions of unseen parents generate debris
clouds, which then shear out and become part of an overall
patchy background of faint material; the particles comprising
the background are continually re-accreting onto the parents,
only to be thrown off in some later collision. According to
this model, about 102 clouds (with 'K'~10-3-10-'» and "lO^kmxlO"*
km size) should be present at any typical time. With a steep
size distribution of parents, even the F ring itself could be
the outcome of a collision between two of the largest
moonlets, in which case the F ring is a temporary feature of
the ring system. We suggest that arcs like those in
52
Neptune^s retinue may be produced by this scheme^ Kolvoord
(7) has developed a perturbation routine to study the orbital
consequences of an embedded moonlet on a narrow ring and is
presently incorporating an exterior satellite perturber into
his program as well as considering the consequences of
particle collisions within the narrow rings.
Showalter and co-workers (3) have interpreted quasi-
periodic optical depth variations that were found on either
side of the Encke gap as the gravitational "wakes" of
a moonlet orbiting near the gap's center and at a longitude
of 32° . A single moonlet of '"10 km radius is able to produce
the observed features. Studies of the signal scatter of the
PPS occultation data by Showalter and Nicholson (8) ,
initiated under this grant, give promise of providing an
independent measure of the size distribution of ring
particles .
bein
prim
atmo
tran
edge
not
4.57
dist
that
The o
g stud
arily
spheri
sition
s seem
procee
Ri^ mu
ributi
reson
rigin and fate of Uranian ring dust is presently
ied (6). Such ring particles have brief lifetimes
due to sputtering and orbital collapse under
c drag. An unresolved quandary is that abrupt
s in ring brightness observed at interior ringlet
to imply that the expected orbital evolution does
The dust discovered on ring plane crossing at
have nearby sources; its broad and asymmetric
about the planet's equatorial plane suggests
electromagnetic forces are at work (5).
d.
St
on
ant
Our studies of faint rings are pertinent to the possible
presence of fine material near Neptune; such grains could
prove hazardous to Voyager 2 as it swings through the ring
plane on its way toward a flyby of Triton; we have been
helping to advise the Voyager project office on the potential
danger to this precious spacecraft.
Several review articles were also prepared during the
funding period. They concerned dust motion (9)j planetary
rings (10) J satellites (11), and satellite orbital evolution
(12).
53
1. L. Schaffer and J. A. Burns (1986). The dynamics of
weakly-charged dust: Motion through Jupiter's gravita-
tional and magnetic fields. Jnl .Geophys . Res . 92, in
press. Abstracts in BAAS 17, 921; BAAS 18, 777-778 and
838, Nat'l Congress Applied Mech . Paper F3a (Austin,
June 1986).
2. M.R. Showalter, J. A. Burns, J.N. Cuzzi and
J.B. Pollack (1986). Jupiter's ring system: New
results on structure and particle properties. Icarus
69, in press.
3. M.R. Showalter, J.N. Cuzzi, E.A. Marouf and
L.W. Esposito (1986). Satellite "wakes" and the orbit
of the Encke gap moonlet. Icarus 66, 297-323.
4. J.N. Cuzzi and J. A. Burns (1986). Charged particle
depletion surrounding Saturn's F Ring: Evidence for a
moonlet belt. Icarus , submitted; Abstracts in BAAS 17,
922; 18, 768; EOS 67, 1077.
5. D.A. Gurnett, W.S. Kurth, F.L. Scarf, J. A. Burns,
J.N. Cuzzi and E. Grun (1986). An analysis of
micron-sized particle impacts detected near Uranus by
Voyager 2. Jnl . Geophys . Res . , in preparation; Abstract
in EOS 67, 340.
6. J. A. Burns, L.E. Schaffer, J.N. Cuzzi, and
D.A. Gurnett (1986). Dust in the Uranian system: Its
origin and fate. BAAS 18, 770-771.
7. R.A. Kolvoord (1986). The effect of an embedded
satellite on narrow rings using a perturbation
approach. BAAS 18, 771,
8. M.R. Showalter and P.D, Nicholson (1986). Saturn's
rings through a microscope: Constraints on particle
size from the Voyager PPS scan. BAAS 18, 767.
9. J. A. Burns (1986). The motion of interplanetary
dust. In Evolution of the Small Bodies in the Solar
System (M. Fulchignoni and L. Kresak, Eds.), in press.
10. J. A. Burns (1986). Rings around planets. In
Evolution of the Small Bodies in the Solar System
(M. Fulchignoni and L. Kresak, Eds.), in press.
11. J. A. Burns (1986). Some background about
satellites. In Satellites (J. A. Burns and
M.S. Matthews, Eds.), University of Arizona Press,
1-38.
54
SUBCENTMETER-SIZE PARTICLE DISTRIBUTION FUNCTIONS IN PLANETARY RINGS
FROM VOYAGER RADIO AND PHOTOPOLARIMETER OCCULTATION DATA
Howard A. Zebker^^ Jet Propulsion Laboratory, G. Leonard Tyler and Essam A.
Marouf, Stanford University.
Analysis of measurements of the scattered and direct components of Voyager 1
radio occultation signals at 3.6- and 13-cm wavelengths yields estimates of the
distribution functions of supracentimeter-size particles and thickness of relatively
broad (>1000 km wide) regions in Saturn's rings (Tyler at al.,1983; Zebker et al., 1985;
Zebker and Tyler, 1984). The amplitude of these signals is mostly unaffected
by particles smaller than a wavelength in size, thus the distribution of these
particles in the rings cannot be determined by analysis of the amplitude data
alone. If, however, measurements of signal amplitude at a shorter wavelength
are combined with the previously analyzed data, we can constrain the shape
of the distribution functions characterizing the smaller particles. In particular,
the additional measurement of amplitude at 0.26 /cm wavelength from Voyager
photopolarimeter occultation observations (Esposito et al., 1983) allows inference
of subcentimeter-size distribution functions for a number of relatively narrow
(30-100 km wide) embedded ringlets in Saturn's ring C and the Cassini division.
These data are available from the January, 1986 Uranian ring occultation also,
and this same technique can be used to estimate corresponding distributions for
the Uranian ringlets.
If we consider size distributions of arbitrary form, many solutions are found that
are consistent with the three available observations of signal amplitude. In order
to limit the formal solution set to functions that are likely on a geophysical basis,
we constrain the solutions to be of the power-law form with sharp lower- and
upper-size cutoffs. We calculate the best-fit power law (in the least-square-error
sense) to the three observations at three wavelengths for several of the embedded
Saturn ringlets— the results are tabulated in Table 1. We note that in each case but
two the inferred power law index is approximately 3, which is similar to the power
law that describes the distribution of supracentimeter particles in the broad ring
features. Thus, it is likely that accumulations of particles in the embedded features
are distributed similary to the particles in the rings as a whole, and that the forces
responsible for the creation and maintenance of of the embedded features are not
highly size-selective in nature.
Mie scattering theory predicts that the measured phase of the radio occultation signal
is highly sensitive to particles ranging from 0.1 to 1.0 wavelengths in size, thus
additional constraints on the subcentimeter-size distribution functions for both
the Saturn and Uranus rings can in principle be derived from radio phase
measurements. However, the observed phase and amplitude data from both sets
of occultations cannot be reconciled with classical Mie theory. Discrepancies of up
to a factor of three are found between predicted and measured values for nearly
all physically-likely distribution functions. We have attempted to account for
these differences by introducing terms in the Mie equations representing proximity
effects of particles (see Zebker et al., 1985) and also for non-sphericity of the particles,
however a reasonable match between theory and observation has not been found.
We are beginning to investigate new models of scattering by dense particulate media
in order to understand the limitations of the Mie approach, and this area of
investigation remains as future work.
55
References:
Tyler, G.L., E.A. Marouf, R.A. Simpson, H.A. Zebker, and V.R. Eshleman, The
microwave opacity of Saturn's rings at wavelengths of 3.6 and 13 cm from
Voyager 1 radio occupation, ICARUS 54, 160-188, 1983.
Esposito, L.W., D.L. Coffeen, C.W. Hord, A.L. Lane, M. O'Callaghan, R.B. Pomphrey,
M. Sato, K.E. Simmons, and R.A. West, Voyager photopolarimeter stellar occultation
of Saturn's rings, J. GEOPHYS. RES. 88, 8643-49, 1983.
Zebker, H.A., and G.L.Tyler, Thickness of Saturn's rings inferred from Voyager 1
observation of microwave scatter, SCIENCE 223, 396-98, 1984.
Zebker, H.A., E.A. Marouf, and G.L. Tyler, Saturn's rings: particle size distributions
for thin-layer models, ICARUS 64, 531-548,1985.
Table 1. Saturn Ringlet Distribution Functions
Ringlet
location
Power law
Lower size
Upper size
km from
Saturn center
index
cutoff, m
cutoff, m
Start
Stop
79230
79260
3.7
0.00316
0.316
84780
84930
2.7
0.00005
0.100
85670
85695
3.6
0.00316
0.178
85930
85970
2.8
0.00032
0.100
86400
86580
2.8
0.00018
0.100
87185
87210
3.0
0.00003
0.178
87300
87325
3.2
0.00178
0.178
87500
87580
3.2
0.00002
3.162
88380
88723
3.1
0.00003
10.00
89800
89930
3.3
0.00178
0.316
90420
90600
2.9
0.00006
10.00
56
Observational Studies of Saturn's Rings
Carolyn C. Porco, Lunar and Planetary Lab/Department of Planetary Science, University of Arizona,
Tucson, Az. 85721.
The objective of this work is to investigate several noteworthy pheonomena in Saturn's rings which
have until now received an inadequate amount of attention. Among these are the periodic variation of
the 'spokes' in the B ring and eccentric features throughout the rings.
One of the major discoveries by Voyager has been the existence of eccentric features within the
predominantly circular rings of Saturn (Smith et al. 1981,1982). Several of these nonaxisymmetric
features are narrow elliptical rings which share many characteristics with the rings of Uranus (Porco
1983; Porco et al. 1984a). In recent work, two new narrow ringlets have been added to the list of
eccentric features in the rings of Saturn (Porco and Nicholson 1986). The ringlet at 1.95i2g sits in the
Huygens gap outside the outer B ring edge and together they comprise an interesting dynamical system.
The B ring edge, also a major eccentric feature which clearly owes part of its shape and kinematics to the
nearby 2:1 inner Lindblad resonance with the satellite Mimas (Porco et al. 1984b), is not completely
described by a simple m — 2 radial distortion expected from this resonance. The 'Huygens' ringlet
behaves, to first order but not entirely, as do the majority of m = 1 narrow ringlets in the solar system
precessing around an oblate plcinet (Fig. la). In addition, it does not exhibit the positive linear width-
radius relation found for many narrow ringlets like the e, a, and p rings of Uranus (Fig. lb). Voyager
Imaging and occulation (RSS, PPS, UVS) data are now in hand, as well as image-processing software
which allows accurate absolute positional measurements to be made in Voyager imaging data. Work
is in progress to re-examine this region of Saturn's rings and to study the possibility of a dynamical
interaction between the outer B ring edge, the Huygens ringlet, and the nearby Mimcis 2:1 resonance.
An understanding of the kinematics and dynamics of this region promises to yield important clues to a
matter of great interest in both theoretical and observational ring studies: the behavior of ring particles
in regions of high optical depth like the outer B ring.
Ever since the Voyager discovery of the broad-band, impulsive Saturn Electrostatic Discharges
(SED) whose spectrum is akin to that of lightning, it was speculated that the spokes observed in
Saturn's B ring might be the visible manifestation of the SED. It has been known for some time that
the variation of spoke activity in Saturn's B ring is modulated at a period equal to 641 ± 5 minutes,
within observational uncertainty equal to that of Saturn's magnetic field, 639.4 ± .05 minutes (Porco
and Danielson 1984). However, the SED's have modulation period of 610 ± 5 minutes (Evans et al.
1981), significantly different that the magnetic field period. There is a strong suggestion of a peak at
a period of 610 minutes in the spectrum of spoke activity (Fig. 2) and current work is focussed on the
investigation of this feature (Porco and Haemmerle 1987) and the problem of the physical connection,
if any, between spokes and the SED.
Evans, D. R., et al. (1981). Nature 292, 716-718.
Porco, C. C. (1983). Ph. D. dissertation, California Institute of Technology
Porco, C. C. and Danielson, G. E. (1984). The kinematics of spokes. Proceedings of the lAU Colloquium
# 75, Planetary Rings, A. Brahic, Ed., ONES, Toulouse, Prance.
Porco, C, et al. (1984a). Icarus 60, 1-16.
Porco, C, et al. (1984b). Icarus 60, 17-28.
Porco, C. C. & Nicholson, P. D. (1986). Icarus , Submitted.
Porco, C. C. & Haemmerle, V. R. (1987). In preparation.
Smith, B. A., et al. (1981). Science 212, 163-191.
Smith, B. A., et al. (1982). Science 215, 504-537.
57
J20 180 140
SOL LONGITUK. dq
100 360
120-
100-
i"
m
40
20
"r-T—T— 1 — r-T — r-T—r—T — r
+
^s
117 75* 'in' 78* 'm!
CtI inUt' 117:84
RADIUS U V? km)
RSS
I I I I I I I
VlT^
117.90
I
.i4iii^.^ M^"i^i
KRIOD (HINUTES)
risurs fren Porco and Danlelson (1984)
58
The Production of "Braids" in Saturn's F Ring
Jack J. Lissauer and Stanton J. Peale (UCSB),
Two models are presented to explain the "braided" structure observed
In some Voyager images of Saturn's F ring. A braided pattern can
be produced from an initially uniform, circular, narrow ring, which
orbits past a nearby satellite (resulting in a wavy pattern) and then
is doubled back upon itself due to gravitational accelerations from an
embedded moonlet (Figure 1). Viewed from the frame which rotates
with the moonlet's orbital angular velocity, the trajectories of the ring
particles traverse one end of the now classic horseshoe orbit. At least
one (and probably both) of the F ring shepherding satellites is mas-
sive enough to produce the required wavy pattern; an icy moonlet 5-10
km in radius would be massive enough to subsequently cause the ring
particles to horseshoe into irregular, short braided strands. Alterna-
tively, if the F ring is composed of two well-separated strands prior to
the passage of the shepherding moon, then the diifering wavelengths
of the wavy patterns induced in the two strands and the subsequent
drift in relative phase can lead to a regular, lengthy braided pattern,
without the need for an embedded moonlet. (Figure 2).
40
20
-
S -20 -
I -40
t-4
.vz
I I I I I I I
68
66
64
62 (degrees)
Fig. 1: Single strand Shepherd plus moonlet perturbations.
n I I I I I I I u I I I I 1 I I I I I I I I I I I I I :
2B0
240
s
(0
220
1
200
180
,«****<...
'•>
I I I I I I I I I I I f I I I I I I I I I I I I i-L' ''
-22 -24 -26 -28 -30 -32 -34(degreRSI
Fig. 2: Two strands, shepherding perturbations only.
59
Experimental Studies On the Impact Properties of Water Ice
F.G. Bridges, D.N.C. Lin, and A.P. Hatzes
Physics Dept. and Board of Studies in Astronomy and Astrophysics,UCSC
We have continued our experimental studies on the impact of ice peirticles at very low ve-
locities. These measurements have applications in the dynamics of Saturn's rings. Initially data was
obtained on the coefficient of restitution for ice spheres of one radius of curvature. The type of measure-
ments have now been expanded to include restitution data for balls with a variety of surfaces as well as
sticking forces between ice particles. We have made significant improvements to this experiment, the
most important being the construction of a new apparatus. Previous mejisurements were made using
iceballs mounted on a disk pendulum which could be made to oscillate for very long periods (40-60
sees). This pendulum was placed in a styrofoam cryostat and cooled using liquid nitrogen. Although
this setup was adequate for performing the first measurements, it suffered from several drawbacks.
First of all the the cryostat used was not completely air-tight. Not only did this preclude the taking
of measurements under vacuum conditions, but atmospheric water vapor was able to leak into the ap-
paratus, thus making it very difficult to make measurements with iceballs that were completely free
from any frost on the impact surface. Secondly, this cryostat was only able to reach temperatures of
around 170° K and only for short periods of time. Measurements of the velocity were made by bouncing
a laser beam off a mirror mounted on the axis of the pendulum. This reflected beam would sweep past
a photocell device prior to and after each collision. The time required for the beam to sweep pcist the
cell was a measure of the velocity of the icebcill. This detector enabled us to mesisure the coefficient
of restitution accurately down to a velocity of 0.02 - 0.05 cm/sec. Reliable data on the coefficient of
restitution is needed for velocities below this.
The new apparatus consists of a smaller version of the disk pendulum and a stainless steel,
double-waUed cryostat. The new apparatus has proved to be a significant improvement over the old
one. Measurements can now be made at temperatures near QCK, comparable to the temperature of the
environment of Satxim's rings, and with much greater temperature stability. With the recent acquisition
of a diffusion pump, the iceball chamber can now be evacuated to a pressure of 10~^ torr. For the actual
measurement of the velocity a capacitive displacement device (CDD) is now used. This device consists
of a set of parallel plates mounted near the top of the disk pendulum. These plates rest between a
similar set mounted on the cryostat. As the pendulum oscillates the plates on the pendulum swing
between those on the cryostat. This device then mesisures the varying capacitance of the plate system
and converts this to a voltage as a function of time which is directly related to the displacement of the
pendulum. The sensitivity of the CDD enables us to obtain accurate measurements of the coefficient
of restititution for velocities near 0.005 cm/sec, even for moderately short periods of oscillation of the
pendulum (10-15 sees). The CDD also has an advantage over the photocell detector in its ability to
measure the displacement of the pendulum as a function of time during the entire collision. This device
not only improves the accuracy of our measurements but makes it possible to obtain data on the contact
time of the collision.
With the old appciratus ice spheres were mjide by freezing water in tennis ball molds. This
had the disadvantage that variations in the shape of the contact surface could not be controlled. We now
use a precision aluminum mold for producing ice spheres with a radius of curvature of 2.5 cm. We also
have additional molds for freezing different radii of curvature on the ice spheres. Figure 1 shows typical
data on the coefficient of restitution for smooth, frost-free ice spheres with different radii of curvature
(2.5, 5, 10, and 20 cm) and taken at the same temperature (123K°). The most obvious feature of these
figures is that the dependence of the coefficient of restitution on the radius of curvature In the range
of .05-1.0 cm/sec is quite weak. As one goes to higher velocities (> Icm/sec) the differences in the
coefficient of restitution between the various ice spheres becomes more noticeable, with spheres with
the larger radius of curvature being more elastic. In the range of 0.1-1.0 cm/sec the data for all four ice
60
spheres can be fit by either a linear or an exponential function with the exponential form providing the
slightly better fit. Data taken at velocities greater than 2 cm/sec should better distinguish between the
two laws, and these measurements are currently in progress. Using a function of the form e(u) = Ce"''"
to fit the data in figure 1, one finds that there is very little vsuriation in C among the different iceballs
(C~ 0.9); however there is a noticeable trend of decreasing 7 with larger radius of curvature. The
behavior of 7 versus R, the radius of curvature is shown as an insert in figure 1. It is a linear relation
of the form 7(i2) = 0.41 — O.IR. For the 5, 10, and 20 cm radii iceballs only the exponential fits to the
data are shown.
A large effect on the coefficient of restitution can be caused by the condition of the contact
surface before the collision. We find that a roughened contact surface or the presence of frost can cause
a much larger change in the restitution measure than the geometrical effect of the radius of curvature.
Figure 2 shows data taken during one run using an ice sphere that had a radius of curvature of 20 cm.
Circles represent data taken while the iceball had a relatively smooth surface. Triangles represent data
taken using the same ball after the cryostat had been evacuated for several minutes at a relatively high
temperature (T=210°if). This allowed the ball to sublimate and thus created a roughened surface on
the ball. Crosses represent data taken after water vapor wjis blown across the surface of the same iceball
thus allowing frost to form. An even further reduction in the coefficient of restitution is evident. At a
velocity of 0.5 cm/sec the sublimated iceball was 20% more inelastic than the smooth ball at the same
velocity while the frosted ball was 30% more inelastic. This is much larger than the variation in the
coefficient of restitution at a given velocity between different iceballs with the same radius of curvature,
which can be as large as 10%.
Work is now in progress to further quantify the effects of sublimation and frosting. We
have recently completed a ga^ handling system that enables us to deposit carbon dioxide, methane, or
ammonia on the surface of the iceball. Effects on the restitution measure with these substances present
will also be examined.
61
o
o
u
o
■<->
t/l
K
o
o
.2 —
1 — I — r
I I r
T=123 K
Ri»n=20cm
Ri»n=10cm _
Ri»n=5cm
Rb«u=2.5cin
=T^
1 1 1 1 1
' 1 '
1 1 1 1 1 1 1 1
1 1 r-'i r=
s-Ci.
-
r
~ f^
—
:
~
E-
r(R)=
0,41-
-O.OIR ~ '
9 — -— =1
z~
1
-=
= 1 1
1 1 1 r 1
1 1 1
11 1 1 [ 1 1 1
!,,.,=
10
15
20
25
J 1 L
J 1__L
J L
Radius pt Curvature (cmV i
J L
.6 .8 1 1.2
Velocity (cm/sec)
1.4
1.6
1.8
Fig. 1j Typical data taken with a smooth iceball of radius 2.5cm (circles). Lines represent exponential
fits to the data of the form e(t;) = Ce~''^ . Fits to the data for balls with larger radii of curvature are
also shown. Insert shows behavior of 7 with radius of iceball.
T—r
1 1 r
-rnr
n I r
T
T— r
-| — r
OA
a^p9
Rban=20cm
T=123''C
O Smooth iceball
A Sublimated iceball
+ Frosted iceball
I II
1
1
I I I
J L
\ I I
J L
J I L
.4 .6 .8 1 1.2
Velocity (cm/sec)
1.4
1.6
l.B
Fig. 2'. Restitution data for an iceball with 20cm radius of curvature. This shows the changes in the
coefficient of restitution after the ball is sublimated (triangles) and then frosted (circles).
62
CHAPTER 3
COMETS AND ASTEROIDS
TSMffiMAIURES MB MINIMJM TIICINSSS OF IHE INACTIVE SUWA(S. LAIEl OF COIET
liLLEY
F. P. Fanale sad J. R. Salvail, Planetary Geosciences Div. , Hawaii Inst,
of GeopliysiGS^ Uniy, of Hawaii, Honolulu^ HI.
The effects of a nonvolatile mantle on the thermal state of a comet
nucleus are investigated. Owe original computer model (Fanale and Sal vail,
1984) was modified so that temperatures can be computed through a thin dust
mantle to the center of a 5 km spherical nucleus in the orbit of P/Hall^.
No attempt is made to simulate the formation of a mantle. Results are
obtained for various specified values of initial mantle thickness and ther-
mal conductivity to determine their effects on temperature profiles through
the mantle. The minimum thickness of mantle that can withstand ejection by
sublimating gasses is also calculated as a function of mantle thermal con-
ductivity. This is assnned to occur when the vapor pressure at the ice
interface exceeds the lithostatic pressure of the mantle. Calculations
were performed for ten or more orbits until temperatures in the mantle
reached a near steaify state. Results indicate that mantles as thin as 4 cm
and 14 cm» for thermal conductivities of 600 and 6000 ergs/cm-s-K, respec-
tively, will remain intact. Surface temperatures as high as 511K at per-
ihelion and 400K at the position of spacecraft encounter were computed at
latitude for an upright, rotating nucleus. Ice interface temperatures
were raised by different amounts during each orbit, depending on mantle
thickness and thermal conductivity, until steady state was reached. After
temperatures in the mantle reached a steady state, ice surface temperatures
were constant throughout the orbit due to the large difference in the ther-
mal conductivities of the mantle and the more compact icy nucleus. These
results imply that relatively small nonvolatile masses emplaced randomly in
comet nuclei could produce an irregular, permanently mantled surface and
could also account for the apparently random location of active areas.
Reference
Faaele. F.P. and J.R. Salvail (1984). An idealized short period comet
model: Surface insolation, H^^O flux, dust flux and mantle evolution.
Icarus 60, 476-511.
65
COMET THERMAL MODELING
Paul R, Weissman, Jet Propulsion Laboratory, Pasadena, CA 91109, and Hugh
H, Kieffer, U. S« Geological Survey, Flagstaff, AZ 86001
The past year has been one of tremendous activity because of the appearance
of Halley's Comet. Observations of the comet have been collected from a
number of sources and compared with the detailed predictions of the comet
thermal modeling program. Spacecraft observations of key physical parame-
ters for the cometary nucleus (size, albedo, dust-to-gas ratio, etc.) have
been incorporated into the thermal model and new cases run. These results
have led to a much better understanding of physical processes on the nucleus
and have pointed the way for further improvements to the modeling program.
A new model for the large-scale structure of cometary nuclei was proposed
in which comets were envisioned as loosely bound agglomerations of smaller
icy planetesimals, essentially a rubble pile of primordial dirty snowballs.
In addition, a study of the physical history of comets was begun, concentra-
ting on processes during formation and in the Oort cloud which would alter
the volatile and non-volatile materials in cometary nuclei from their pris-
tine state before formation. Dr. Gary Herman of Tel Aviv University spent
one year at JPL as a NRC post-doc working on two interesting research
tasks: internal temperatures in icy nuclei, and radiative transfer in dusty
cometary comae.
The thermal modeling of Halley's Comet has shown that the asymmetric behav-
ior of Halley's light curve pre- and post-perihelion cannot be explained by
heat flow into sub-surface layers on the inbound leg of the orbit, providing
an additional energy source as the comet moves away from perihelion. Within
2 AU of the sun, high values of surface thermal conductivity can yield post-
perihelion brightenings of only about 10 or 20%, as compared to the 100 to
200% brightenings that are actually observed. Previous models which sug-
gested this behavior did not allow for radiative cooling of the cometary
nucleus at night.
The correct explanation for post-perihelion brightening appears to be sea-
sonal changes on the inclined, rotating nucleus. As a result of the comet's
highly eccentric orbit, as it rounds perihelion there is a very abrupt
change in the declination of the sub-solar point from the southern to
northern hemispheres of the nucleus. On the way towards perihelion the
northern hemisphere receives only modest heating, in fact, none around the
circumpolar region. The sudden change post-perihelion, at a time when the
comet is very close to the sun, causes rapid temperature increases and
resulting thermal stresses. The compressional hoop stresses on the non-
volatile cometary crust material causes cracking and strike-slip fractures.
Plates of crustal material are broken loose from the nucleus and substantial
new areas of fresh ice are exposed beneath the crust. The activity continues
to build as the sun moves northward in declination following perihelion.
Eventually, the decreasing solar insolation as the comet moves away from
the sun causes the activity to subside.
Activity is less on the inbound leg because of two factors « First of all^
the heating is more gradual and the thermal stresses can be accomodated
more readily without catastrophic failure of the non-volatile crustal
material. Second, the gradual heating likely depletes near surface layers
of volatiles while building a thicker non-volatile crust layer over the ice.
Detailed solutions for the seasonal behavior of the Halley nucleus are sens-
itive to rotation pole orientation^ However, all the suggested pole orien-
tations for Halley are within about 40 degrees of each other, with suggested
obliquities of 20 to 30 degrees. Other factors such as the nucleus rotation
period, a still poorly determined parameter, and the triaxial spheroid
shape of the nucleus also will affect the detailed gas production rates
that are derived from the comet model as a function of orbital position.
Comparison of the comet thermal model results with the observed behavior of
the Halley nucleus versus heliocentric distance showed that the fraction of
active sublimating area on the nucleus surface was not constant throughout
the orbit but changed in unpredictable ways (though consistent with the
seasonal dependence explanation above) . This is further proof of the
heterogeneous nature of the nucleus and of cometary phenomena in general.
The surface heat flow becomes important with regard to the behavior of the
nucleus at large heliocentric distances where the energy going into heat
flow is comparable to that going into sublimation. For high heat flows the
comet does not "turn on" until relatively close to the sun, while for low
heat flow the coma can become visible at over 6 AU from the sun. Given the
observed turn on of the Halley coma at 5.8 AU inbound in early 1985, one
can set an approximate value for the thermal conductivity of about one-tenth
that of solid crystalline water ice. This is a relatively low conductivity,
though not as low as observed for some dusty satellite regoliths in the
solar system.
The proposed new model for cometary nuclei, known as the "primordial rubble
pile," considered what a cometary nucleus should look like based on present
scenarios for planetesimal formation in the outer solar system, and attempt-
ed to explain a variety of observed cometary phenomena. The lack of major
energy sources means that cometary material will likely not be brought to-
gether into a single, well consolidated body, but will retain the composite
structure of an agglomeration of smaller dirty ice snowballs. Phenomena
such as cometary outbursts and splitting might be explained by such a
structure, as smaller pieces break off to become secondary nuclei, and
freshly exposed faces result in sudden brightening and activity from the
main nucleus. Evidence from radar studies of large debris clouds around
nuclei tends to support this suggestion of a possible "rubble pile" struc-
ture, with small debris being briefly elevated off the nucleus surface and
then falling back very slowly in the weak cometary gravity field.
Imaging of the Halley nucleus did not have sufficient resolution to determine
if the primordial rubble pile model is correct, Giotto images did show a
highly irregular nucleus with surface roughness on the order of 10% the mean
radius, or more. But better images, presumably from the GRAF mission, will
67
be needed to resolve this question. In addition, it is possible that
after long years of physical evolution , the original rubble pile structure
of the nucleus is hidden by external modification and mass movement.
Dr« Herman developed a detailed analytical solution to the nucleus heating
problem and a one-dimensional numerical model of the means by which cold
nuclei from the Oort cloud warm as both their perihelia and orbital semimajor
axes decrease* He showed that previous models were inaccurate because they
ignored surface heat flow. The internal temperature is a complex function
of both the comet's semimajor axis and eccentricity j as well as the nature
of the ice making up the nucleus, i.e., amorphous versus crystalline. The
time required for the nucleus to reach equilibriimi temperature is often
quite long, exceeding the dynamic time scale for substantial changes in
the comet's orbit due to close planetary encounters.
Another research effort was concerned with radiative transfer in dusty
cometary atmospheres. It had been shown that a dusty coma can increase the
total energy reaching the cometary nucleus as a result of multiple scattering
and thermal emission by the dust. Dr. Herman showed that most past estimates
of this phenomena tended to over-estimate the effect, and that it was diffi-
cult to get a large energy increase* However, the coma does serve to
redistribute energy around the cometary nucleus, illuminating the night
hemisphere while cutting down the total radiation reaching the dayside.
This would lead to a more isothermal nucleus when the coma opacity was
high. A Monte Carlo simulation of the radiative transfer by Dr« Herman
and Dr» H. Salo showed that a modest, on the order of 10%, increase in
total flux reaching the nucleus was possible, if one assumed forward scat-
tering particles (g = 0»7), and an accelerating dust velocity due to entrain-
ment in the evolving gas«
Finally, an analysis of possible mechanisms for modifying cometary nuclei
during their formation stages, and/or during their long residence time in
the Oort cloud was begun» Because of the small size of the nuclei and their
formation in zones far from the sun where orbital velocities are small, their
total gravitational potential energy and degree of compaction is probably
quite small. The radii of the cometary nuclei are sufficiently small that
they were likely not substantially heated by long-lived radio-isotopes.
Short-lived isotopes like aluminum 26 may have melted the interiors of
comets, but only if comets formed over a time span short as compared to the
lifetime of that radionuclide* That appears to be unlikely. Sputtering of
nucleus surfaces by galactic cosmic rays is an important process, removing
much of the volatiles and polymerizing all the carbon compounds down to a
depth of a meter or more. This leads to the interesting possibility that
comets may have already developed non-volatile crustal layers before they
entered the planetary system. Although comets might accrete a thin veneer
of interstellar dust and gas while resident in the Oort cloud, it appears
more likely that hypervelocity impacts by interstellar dust grains result
in a net erosion of the cometary surfaces, though the effect here is only
on the order of centimeters, as compared with modification to a depth of a
meter or more by cosmic rays. Thus, comets may not be as entirely pristine
as originally thought, and it will be important to consider these various
modification processes in interpreting the results form the Halley's Comet
missions, and from future missions such as GRAF and CNSR,
68
IHE LOSS AND DEPTM OF CO^ ICE IN £»MT NUCLEI
F. F. Fanale aad J. R. SalTail^ Planetary Geosciensss Div. s Hawaii Inst,
of GeopkysisSs Univ» of Hawaii^ lonolBlu^ HI.
An analytical model has been developed to simulate the material dif-
ferentiation of a cometary nucleus composed of water isej putative
unclathrated C0„ ise and silicate dust in specified proportions. Selective
sublimation of any free CX>- ice present in a new comet would produce a sur-
face layer of water ice ana dust overlying the original W„ rich material.
This surface layer reduces the temperature of buried CO. ice and restricts
the outflow of gaseous CX)^, On each orbit, water sublimation at analler
heliocentric distances temporarily reduces the thickness of the water ice
and dust layer and liberates dust. Most of the dust is blown off the
nucleus, but a small amount of residual dust remains on the surface. (cf.
Houpis et al., 1985) Our model includes the effects of nucleus rotation*
arbitrary orientation of the rotation axis, latitude, heat conduction into
the deep interior of the nucleus and restriction of CX>- gas outflow by the
water ice and dust layer, features that were not included in the Houpis et
al. model. Specifically, we investigate the effects of the permeability of
the surface water ice layer, the nucleus rotation rate and the latitude.
The loss rate of CX)„ and the resultant depth of CX)- ice are shown to be
most sensitive to the permeability of the water ice and dust layer. For a
hcmogeneous, initially unmantled comet placed in the orbit of comet Hall^,
it is shown that the CO. ice attains a steady state or cyclic relationship
between CX)^ depth and orbital position within several revolutions. If the
nucleus contains 25% by mass of CX)„, our results indicate that CX)., ice is
always within several meters from the surface at airp^ location for a nucleus
of low obliquity and that CX)^ iee is nearest to the surface at the equator
shortly after perihelion. Under these conditions the sublimation of CO-
ic® is always significant and becomes the dominant gaseous species beyond
4MI« This result is probably generally valid for unmantled portions of
most comets and qualitatively simulates the behavior of an abundant, highly
volatile component in an H-O/sil icate matrix. Comparison of these and simi-
lar results with observations could yield information regarding the permea-
bility and chemical composition of cometary material and suggest sampling
strategies to minimize f raotiosmtion effects.
Reference
Moupis, H.L.F. , W.H. Ip and D.A. Mendis (1986). The chemical differentia-
tion of the cometary nucleus: The process and its consequences. Astrophy-
sical J., in press.
DYNAMICS OF LONG PERIOD COMETS
Paul R. Weissman, Earth and Space Sciences Division, Jet Propulsion
Laboratory, Pasadena, CA 91109 USA
The appearance of Halley's Comet in 1985-86 and the related emphasis on
research on physical models of cometary nuclei, led to a more moderate
pace for the dynamical studies of the Oort cloud and the motion of long-
period comets this year. Specific areas studied included the dynamical
evolution of cometary showers as a result of stars passages through the
inner Oort cloud and the possible relationship to observed stepwise mass
extinctions at geological boundaries, revised estimates for the total
mass of comets in the Oort cloud as a result of lessons learned from the
spacecraft encounters with Halley's Comet, and study of the possible dy-
namical sources for the short-period comets in the solar system as part
of a wider study of physical processing of cometary nuclei prior to their
becoming visible comets.
The work on cometary showers used a Monte Carlo simulation of the evolu-
tion of cometary orbits under a combination of planetary, nongravitation-
al, and stellar perturbations, and with physical removal by disruption,
sublimation of all volatiles, and collision, A major cometary shower
lasts about 2 to 3 Myr after the initial perturbation, with peak flux
rates of about 9,000 comets crossing the Earth's orbit per year (600 times
the usual flux). The average comet makes 8,5 returns with an average
lifetime of about 0.8 Myr. Such an intense random shower would be caused
by a one solar mass star passing 3,000 AU from the sun, would be expected
perhaps once every 500 Myr, and might produce on the order of 20 signifi-
cant impacts on the Earth.
The actual mechanism for initiating showers was found to be quite interest-
ing. As compared with Hill's (1981) simple loss cone analysis, it was
found that the total number of comets entering the planetary region was
somewhat less, and the fraction perturbed to small perihelia (Earth-cross-
ing orbits) was lower still. In addition, the intensity of showers drop-
ped off sharply with more distant stellar passages, roughly as the inverse
square of the encounter distance. The number of comets peturbed into the
inner solar system is very much a function of the distribution of orbital
semimajor axes in the inner Oort cloud; the figures above reflect an
inner cloud that is only modestly centrally condensed. If the distribu-
tion of orbits in the cloud is steeper then distant encounters would be
less effective at inducing large numbers of comets into cometary showers.
The timescales for cometary showers and that for disruption of terrestrial
environments as revealed in the fossil record at extinction boundaries,
are comparable. This does not prove a causal relationship and may be
entirely coincidental, but it does demonstrate that comet showers may
provide a plausible explanation for some biological extinction events.
The main problem with this hypothesis is that extinctions are roughly 10
times more frequent than the expected rate of major cometary showers.
70
The estimated mass of comets In the Oort cloud has increased dramatically
as a result of the revised estimate for typical cometary albedos based on
the Halley spacecraft encounters. The term "dirty snowball" has tended to
mislead people into thinking that comets were relatively bright objects,
perhaps gray in color. In fact, comets are really "frozen mudballs" with
the low albedo of their non-volatile constituents. While lowering the
albedo causes an increase in the mass estimate, other factors such as
shape, improved population estimates based on new dynamical modeling, and
the fraction of active area on the nucleus, serve to reduce the total mass
estiamte. Using a revised albedo of 0.05 based on the Halley encounters,
it was found that the mass of comets in the outer, classical Oort cloud
was 25 Earth masses, and the mass of the inner Oort cloud which serves as
a reservoir to replenish the outer cloud is perhaps 250 Earth masses. The
latter estimate is highly uncertain because of the lack of detailed dynam-
ical modeling of the inner Oort cloud to date. Also, present attempts
to estimate the density of cometary nuclei based on modeling of nongravi-
tational forces on Halley are still highly uncertain.
The question of the source of the short-period comets is one that has taken
on new meaning becaue of the proposed spacecraft missions to comets. To
interpret the cosmochemical record found in comets it is necessary to know
where they formed, and where they have been since their formation. Halley
results have already shown that the comet appears to have formed from the
same material as the rest of the planetary system. But the dynamical
history of Halley prior to its being peturbed into its present orbit
remains a mystery. The classical view is that short-period comets are
long-period comets from the Oort cloud, captured by Jupiter perturbations.
Recently, it has been shown that the inner Oort cloud should provide a
dynamically more efficient source as comets trickle across Neptune's orbit
and are passed down by planetary perturbations into the inner solar system.
However, Halley 's retrograde orbit argues against this (for Halley only)
because the comets derived from the inner Oort cloud should be in direct
orbits. Because of the chaotic nature of cometary orbit evolution, and
the large number of comets we have to deal with, it is difficult to rule
out any dynamical path for the origin of Halley or any other short-period
comet. Work on this problem will continue.
71
OBSERVATIONS ON THE MAGNITUDE-FREQUENCY DISTRIBUTION
OF EARTH-CROSSING ASTEROIDS
Eugene M, Shoeraakerj U=S. Geological Survey^ Flagstaff j ffi 86001
Carolyn S, Shoemaker, Arizona Research and Technology, Flagstaff j AZ 86002
During the past decade, discovery of Earth-crossing asteroids has
continued at the pace of several per year; the total number of known Earth
crossers reached 70 as of September, 1986 » These objects comprise 36 numbered
and 34 unnumbered asteroids, 11 of which are lost; 6 are Atens, 41 are
Apollos, 22 are Earth-crossing Amors, and one of the Earth crossers has a
present perihelion distance of greater than 1,3 AU, The sample of discovered
Earth crossers has become large enough to provide a fairly strong statistical
basis for calculation of mean probabilities of asteroid collision with the
Earth, the Moon, and Venus, It is also now large enough to begin to address
the more difficult question of the magnitude-frequency distribution and size
distribution of the Earth-crossing asteroids.
Absolute V magnitude, H, has been derived from reported magnitudes for
each Earth crosser on the basis of a standard algorithm that utilizes a
physically realistic phase function (Minor Planet Circular 10193), On
average, H = V(1,0) - 0,3, where VCljO) is the absolute magnitude based on a
linear phase function. Fairly reliable observations of magnitude are
available for most numbered Earth crossers, but observations are sparse for
unnumbered objects and commonly consist only of visual estimates of B made
from photographic plates. In the latter case, the derived value of H is
calculated only to the nearest half magnitude.
The derived values of H range from 12,88 for (1627) Ivar to 21,6 for the
Palomar-Leiden object 6344, which is the faintest and smallest asteroid
discovered. The absolute magnitude of Ivar is close to the magnitude
threshold for completeness of discovery of asteroids in the inner part of the
main asteroid belt. It is possible that one or two more Earth crossers as
bright or brighter than Ivar remain to be discovered, but it is unlikely that
there are as many as ten. Completeness of discovery drops with increasing
magnitude. Shoemaker et al. (1979) estimated the population of Earth crossers
to V(1,0) = 18 (equivalent to H = 17.7) at ~1300; the observed number at
H = 17.7 is 50, which corresponds to an estimated completeness of discovery of
about 4%,
If we adopt the estimate of 1300 for the total number of Earth crossers
brighter than H = 17.7 and assume that the number brighter than H = 12,88 does
not exceed 3 Earth crossers, a plausible magnitude-frequency distribution for
the population to H = 17,7 can be represented approximately by
N = 3.17 e"^'28 (13-H)^ ^^
where N is the cumulative frequency (fig. 1), This distribution corresponds
to a monotonically decreasing completeness of discovery with increasing
magnitude above 13,2, which is expected from circumstances of asteroid
discovery. The distribution represented by eq, (1) is much steeper than the
estimated average magnitude-frequency distribution of main belt asteroids.
72
which can be represented approximately by
N <. e-0'9(-H), (2)
To the extent that Earth-crossing asteroids are derived as collision fragments
from main belt asteroids j the size and magnitude distribution of Earth
crossers can be expected to be steeper than those of main belt asteroids.
Most collision fragments that become Earth-crossing objects probably are
displaced in orbital element phase space into secular resonances by Impulses
of the order of hundreds of meters per second. Small collision fragments tend
to receive larger impulses than large fragments, and small fragments probably
are delivered preferentially to secular resonances and ultimately to Earth-
crossing orbits*
The exponent (size index) of the cumulative size-frequency distribution
of the impactors that produced the post-mare lunar craters >10 km diameter has
been estimated at -1,62 (Shoemaker, 1983). A power function of the diameter
with this exponent corresponds to an exponential magnitude-frequency
distribution for the impacting objects of the form
-l'^2-(-H)
N«e5 1°Se^ ^ ^^-0»75(-H)^ ^3^
under the condition that the range of albedo is similar for objects of
different sizes. This derived magnitude distribution for post-mare Impactors
is much flatter than eq. (1) and also is flatter than the magnitude
distribution of discovered Earth-crossing asteroids In the range 13 < H < 16
(fig. 1). If 1300 is adopted as the cumulative number to H = 17,7, then the
number of objects predicted by (3) at H = 13 is 30 times higher than the
number observed « Conversely, if discovery of Earth crossers were considered
complete at H = 13, then fewer objects than have already been discovered to
H = 17.7 would be predicted by (3). From the ratio of the frequency of
discovery of new Earth-crossing asteroids to the frequency of accidental
recovery of those already known, it is certain that the number remaining to be
found to H = 17.7 is many times greater than the number now known. If it is
assumed that most 20-km-diameter, post-mare impact craters have been produced
by asteroid impact, then the difference in exponent between (1) and (3)
suggests that no more than 1/3 of the craters with diameters >50 km were
formed by Earth-crossing asteroids. The large lunar craters probably have
been formed chiefly by impact of comets.
References
Shoemaker, E.M., Williams, J.G,, Helln, E.F,,and Wolfe, R.F., 1979, ±n_
Asteroids, ed. T, Gehrels; Tucson, University of Arizona Press,
p. 253-282,
Shoemaker, E.M,, 1983, Ann, Rev, Earth Planet, Scl. 11, p. 461-494.
73
:^
3.S
3.0
>> 2.5
u
«
O*
o
^ 2.0
>
3
o
1.5
0>
O
-' 1.0
O.S
°-^
N.1300
-• • •-
21
20
19
17
H
Figure 1. Magnitude-frequency distribution of Earth-crossing asteroids.
ASTEROID FAMILIES, DYNAMICS, AND ASTROMETRY
J. Q= Hilliams aad_J. Glbsoa
Jet Propulsion Laboratory, Pasadsna, CA, 91109
The proper elements and family assignments for the 1227 Palomar-Leiden
Survey asteroids of high <iuality have been tabulated and included in a
paper which is in press in Icarus. In addition to the large table, there
are also auxiliary tables of Mars crossers and commensurate objects, histo-
grams of the proper element distributions, and a discussion. Probably the
most important part of the discussion describes the Mars crossing boundary,
how the closest distances of approach to Mars and Jupiter are calculated,
and why the observed population of Mars crossers should bombard that planet
episodically rather than uniformly,
A start has been made on classifying families, the debris left over from
collisions between asteroids, based on the relative sizes of the members,
particulary with respect to the largest object, and where the largest
object (s) lies in the family. The taxonomic classes of the members are
also being compared. This undertaking is about 30H complete. Very few
families appear to have resulted from total disruption and dispersal of the
parent body.
Analytical work has been done to derive velocity distributions of family
forming events from proper element distributions subject to assumptions
which may be appropriate for cratering events. The more conventional
approach uses an "average" value for the two orbital angles (true anomaly
and argument of latitude); the family distributions can also be combined
assuming a uniform distribution of the angles to get an overall velocity
distribution. All researchers who have tried to get velocity distributions
have found that the results from the different proper elements differ
radically, with the semimajor axis giving the least and the inclination
giving the largest velocities (by a factor of two or three). It is usual
to dismiss this discrepancy by attributing it to the accuracy of the proper
elements. While this was a plausible explanation 20 years ago,
improvements in the calculation of proper elements and the presence of
small structure in some families (there is a very tight core in family 33
and the inclination spread of family 128 is quite narrow) argue otherwise.
Selection effects may be present in the less well populated families, but
the strong asymmetry is also present in the more populous families. On the
assumption that the effect is real, several possible causes have been
considered, but all seem to be flawed. For example, the passage of weak
secular resonances after family formation would scatter e and sin i without
scattering a, but it would also leave a boxy morphology which is not seen.
It has to be said that this asymmetry is not understood, but it could be an
important problem.
During the past year software has been developed for a microcomputer to
permit plotting the proper elements. Three orthogonal views are generated
and stereo pairs can be printed when desired. This program was created for
the study of asteroid families (a similar capability of long standing on a
mainframe computer was lost when the high resolution graphics capability
75
was removed). In a cooperative effort with Ea T^desco and Q. Veeder this
software has been_aised to examine a sample of the best physical properties
data that they have amassed. This work is presently ongoing, but so far Pv
and two colors permit ten plausible taxonomic classes to be recognised.
The astrometry task is directed toward measuring and reducing positions on
faint comets and the minor planets with less common orbits. The obseva-
tional material is CCD frames taken with the Palomar 1.5m telescope. There
is a need to measure Schmidt plates because the CCD is too small to capture
an array of reference stars with known positions. It is necessary to use
an existing plate to measure the fainter stars on the CCD frame with
respect to the less common catalogue stars. During the past year positions
of 10 comets and IS different asteroids have been published on the Minor
Planet Circulars. Many of these were early or late observations which
particularly strengthen the orbit determination. In summary, asteroid
positions were given on 10 planet crossers (1981 PB, 1981 VA, 1982 HR, 1982
TA, 1985 JA, 1985 PA, 1985 TB, 1985 WA, 1985 KB, and 198S EB), 1 Pallas-
type object (198S AE), and 5 more ordinary minor planets. The comets were
1980 KI (Encke), 1982 i (Halley), 1985 f, h, j, k, and p and 198fi a, b, and
c. These comet and asteroid positions were published on Minor Planet
Circulars 9844, 9845, 9849, 9984, 9993, 9994, 10006, 10092, 10200, 10204,
10224, 1022S, 10272, 10369, 10370, 10459, 10477, 10478, 10487, 10488,
10591, 10592, 10595, 10596, and 10690.
76
1986 DA MB 1986 Mt M-CLASS ASTEROIDS IN NEM-EARM OBBITS
Jonatlian Gradie, Planetary Geoscienees DiYisioa, HIG, 2525 Correa Road,
Honolulttj II 96822 sad Edward Tedeseo, Jet Propulsioii Laboratory^
Pasadena, CA 91109.
The eartli-approasliiag asteroid population is composed of asteroids in
orbits with, lifetimes short eompared with the age of the solar system.
These objects which are comprised of Atea, Apollo, and Amor asteroids must
be replenished frsai either cometary or mainbelt asteroid sources since
lifetimes against collision with or ejection by a planet are on the order
of 10 to 100 million years. Without a source, the number of such objects^
postulated to be as large as 1,500 (Shoemaker, 1979), would be quickly
depleted. Suggestions for sources has ranged from cometary (Opik, 1951) to
a combination of asteroidal and cometary (c.f. Shoemaker, 1979; Wetherill,
1986). A cometary source implies that the compositions consistent with the
non-¥Olatile remnants of a comet (Hartmaan, et al . , 1986) where as an
asteroidal source would imply that the compositions consistent with the
asteroidal source region.
The physical study of Earth-approaching asteroids is constrained by
the generally long period between favorable apparitions and poorly known
orbits. For example, although 88 such asteroids have been discovered
through 1985 only 47 had orbital elements sufficiently reliable to receive
permanent numbers (Hahn, 1986). For this reason, these objects must be
studied as "targets of opportunity" as soon as they are discovered and
before orbital elements are firmly established and a permanent number is
assigned. Such is the case for objects 1986 DA and 1986 BB,
1986 DA and 1986 M were discovered on 16 February 1986 by M. Kizawa,
Shozouka, Japan and on 4 SJarch 1986 by E. Shoemaker and C. Shoemaker,
Palomar Observatory, respectively (I^ Circ. 4181 and 4191). 1986 DA is a
member of the Amor group since its orbital elements cause it to cross the
orbit of Mars but not the Earth whereas 1986 EB is a member of the Aten
group since its orbital elements cause it to cross the orbit of the Earth
even though its semimajor axis is less than 1 AD.
Broadband spectrophotometry on the Johnson UBVR system and the Eight-
Color Asteroid Survey system (Tedesco, et al . , 1982) were obtained at Kitt
Peak Natioiml Observatory and on the Johnson JHK system and at 10 and 20
microns at the NASA Infrared Telesco^ Facility at Mauna Kea Observatory
(Tedesco and Gradi®, 1987). These observations were used to determine the
absolute visual magnitudes and to derive the visual geometric albedos and
dimeters on the IRAS system (Lebofsty, et al , , 1987) given in Table I.
Table
I. IBV colors
and mean albedos and
diameters.
Obj ect
H G
U-B B-¥ pv
Diameter (km)
1986 DA
1986 m
15.94 0.25
15.94 0.25
0.21 0.70 0.14
0.24 0.71 0.19
2.3
2.0
77
The IB? solors and tk© albedos are CBOBgli to imiquely classify these two
objects on the taxonomic system of Gradie aad Tedesco (1982) as class M,
the first of these objects to be found in the planet crossing population^
The spectral reflectance properties and gecmeteric albedos of the M-
class asteroids are consistent with compositions analogous to the iron-
nickel meteorites or the enstatite-metal assemblages of the eastatite chon-
drites (Zellner, 1979). Radar observations by Ostro, et al. (1985) of the
BFclass asteroid 16 Psyche indicate a body nearly entirely of metal. These
results imply that both 1986 DA and 1986 SB are probably nearly entirely
metallic in composition^ perhaps similar in gross composition to the iron
meteorites.
The identification of two objects probably of entirely metallic compo-
sition has important implications for 1) the origin of the objects
currently found in near-earth orbit, 2) the timescales for the delivery of
iron meteorites to the earth and 3) the sources of raw materials for space
industrialization.
Tedesco and Gradi® (1987) examine the issue of the source (s) of the
near-Earth asteroids population by comparing the classifications on the
scheme employed by Gradie and Tedesco (1982) of 38 such asteroids. Those
asteroids for which an unambiguous C, S« or M classification could not be
assigned were called "Others" (e.g. D, E, F, P, R or unclassif iable) . Five
objects were classified as C, twenty-four as S, two as M (1986 DA and 1986
EB), and seven as "Other". The predominance of the well-known C, S, and M
classes of asteroids ( >80%) in this population strongly suggests that the
source region is the asteroid belt unless cometary nuclei are composition-
ally indistinguishable frcan the C, S» and M class of asteroids. Further-
more, the presence of M-olass asteroids in the near-Earth population argues
that the source in the asteroid belt must be close to the 3:1 and 5:2 Kirk-
wood gaps since these are the only regions in the belt where M-class
asteroids are found.
If the source of most of the near-Earth objects is indeed the asteroid
belt as our observations suggest, then a method for removing extinct nuclei
of short period comets must be found since the rate of production of short
period comets fr«m the long period comets is relatively large. We suggest
that the lack of a large number of objects with compositions consistent
with that expected for short period comets, i.e., dark and spectrally
reddened (Hartmann, et al., 1986), argues that either the majority of
comets lack cohesive, volatile-free cores and end up as meteor streams or
the core of a comet may be so friable that it cannot survive intact as long
as asteroidal material.
The two metallic objects , 1986 DA and 1986 E8, provide examples of
near-Earth, intermediate parent bodies of some iron meteorites. Using the
meteorite production model of Greenberg and Chapnan (1983) we calculate
that half of the iron meteorites should come from these two objects. These
irons should have cosmie-ray exposure ages of < 100 million years, the mean
lifetime of most objects in planet-crossing orbits. However, the cosmic-ray
78
exposiixe ages of nearly all iron meteorites measure to date are 4 to 20
times older which implies that these irons came directly from the asteroid
belt as meter-sized objects without the need for an intermediate parent
bodjr. This discrepancy suggests that either that meteorite production from
near-Earth iron objects is extremely inefficient or, as Wasson (1985)
points out, the small nunber of irons with cosnic-ray exposure ages less
than 200 million years my be the result of experimental bias.
Finally, we note that 1986 DA and 1986 EB, if indeed similar in gross
composition to the iron-nickel meteorites, could prove to be valuable
sources of raw materials for space industrialization since relative tran-
sportation costs to and from these objects could be greatly reduced. Such
objects would contain not only iron (90-95 wt%) and nickel (5-10 wt%) but
also cobalt (0.6 wt%) and sizeable trace amounts of other elements includ-
ing gold, platinmn group metals.
REFERENCES
Gradie, J. and Tedesco, E.F. (1982). Science 216 , 1405.
Greenberg, R. and Chapnan, C. R, (1983). Icarus 55 , 455,
Hartmann, W.K. , Tholen, D.J. and Cruikshank, D.P. (1986). Bull . Amer .
Astron . Soc . 18 » 800-801.
Lebofsky, L.A. , Sykes, M, V., Tedesoo, E.F. , Veeder, G.J. Matson, D.L. ,
Brown, R,H. , Gradie, J., Feierberg, M.A. , and Ru<fy, R.J. (1987). Icarus .
in press.
Opik. E.J. (1951). Proc. Ropv . Irish Acad . 54 A . 165.
Ostro, S.J. , Campbell, D.B. and Shapiro, I.I. (1985). Science 229 , 442.
Shoemaker, E.M. , Williams, J.G. , Helin, E.F. , and Wolfe, R.F. (1979). In
Asteroids edited by T. Gehrels (Univer. Arizona Press, Tucson), p. 253.
Tedesco, E.F. and Gradie, J, (1987). Astron . J., submitted.
Tedesco, E.F. , Tholen, D.J, and Zellner, B, H. (1982). Astron . J. 87 .
1585.
Wasson, J. (1985). Meteorites . (W. H. Freeman and Company, New York),
Zellner, B. H, (1979). In Asteroids , edited hy T. Gehrels (Univ. Arizona
Press, Tucson), p. 783,
79
EVOLUTION OF THE INNER ASTEROID BELTi PARADIGMS AND PARADOXES FROM
SPECTRAL STUDIES, Michael J. Gaffey, Department of Geology, West Hall^
Rensselaer Polytechnic Institute^ Troy^ New lork 12180-3590»
Recent years have witnessed a significant increase in the sophistication
of asteroidal surface material characterizations derived from spectral data.
An extensive data base of moderate to high spectral resolution^ visible and
near-infrared (■~0^5~2,5;im) asteroid spectra is now available (1-4),
Interpretive methodologies and calibrations have been developed to determine
phase abundance and composition in olivine-pyroxene assemblages and to
estimate NiFe metal abundance from such spectra (5-7). A modified version of
the asteroid classification system more closely parallels the mineralogic
variations of the major inner belt asteroid types (8), These improvements
permit several general conclusions to be drawn concerning the nature of inner
belt objectsi their historyj and that of the inner solar systemj and the
relationship between the asteroids and meteorites.
Essentially all large inner belt asteroids have - or are fragments of
parent bodies which have ~ undergone strong post-accretionary heating,
varying degrees of melting and magmatic differentiation, and' subsequent
collisional disruption. The surfaces of the dominant S-type asteroids appear
mostly to be the exposed metal-rich internal layers of differentiated or'
partially differentiated parent planetesimals (7,9). The shift from an S-type
dominated inner belt population to a C-type dominated outer belt population
(10) appears to be primarily the result of post-accretionary heating of the
inner belt and not the signature of a radial compositional gradient in the
original solar nebula. The S-type asteroids are predominantly olivine-metal
assemblages with a relatively minor pyroxene component [ol/px>2,5] (3,9), but
exhibit a significant range of variation. These asteroids show a systematic,
but not yet well characterized, mineralogic variation with semi-major axis.
Large S-type family asteroids exhibit greater lightcurve amplitudes in
general than large non-family S-objects, and the mineralogic range among
members of the S-families is much smaller than the range in the background S-
type population. This suggests that the S-type asteroid families represent
relatively recent (<4byr) collisions onto the cores of previously disrupted
parent bodies. A variety of additional constraints (heating requirements for
small bodies, rarity of pure olivine mantle fragments, meteorite heating
ages) suggest that the thermal evolution of the inner belt occurred very
early, followed quickly by disruption during single collision events with an
intruding flux of large planetesimals, similar to those invoked to abort the
growth of Mars (11).
REFERENCES! (1) Chapman, C.R. and Gaffey, M.J. (1979) in Asteroids (T.
Gehrels and M.S. Matthews, Eds.), U. Arizona Press, pp. 655-687. (2) Zellner,
B,, Tholen, D,J», and Tedesco, E.F. (1985) Icarus 61, 355-416. (3) Feierberg,
M.A., Larson, H.P., and Chapman, C.R. (1982) Astrophys. J. 257 , 361-372,
(4) Bell, J.F,, Hawke, B.R., Owensby, P.D,, and Gaffey, M.J, (1985) Bull. Am.
Astron. Soc. 17 , 729. (5) Adams, J.B. (1974) J. Geophys. Res. 79 , 4829-4836.
(6) Cloutis, E.A,, Gaffey, M.J., Jackowski, T.L., and Reed, K.L. (1986) J.
Geophys. Res., submitted, (7) Gaffey, M.J. (1986) Icarus , in press.
(8) Tholen, D.J, (1984) Ph.D. Diss., U, Arizona, Tucson, (9) Gaffey, M.J,
(1984) Icarus 60, 83-114. (10) Gradie, J., and Tedesco, E. (1982) Science 216 ,
1405-1407. (11) Weidenschilling, S,J. (1975) Icarus 26, 361-366.
This research has been supported under NASA Grant NAGW-642,
80
Meteorite Spectroscopy and Characterization of Asteroid Surface Materials
Tiichael J.Cjaffey, Geology Department, KeiTsselaer Polytechnic Institutes
Troy, New York 12180-3590.
The purpose of this research effort is to improve our understanding of
the origiQj evolution, and inter-relationships of the asteroids; of their
relationships to the meteorites; and of the conditions and processes in the
early inner solar system. The surface mineral assemblage and the surface
heterogeneity of selected minor planets is determined from analysis of
telescopic spectra to provide the data base to accomplish these goals.
The asteroids represent the sole surviving, in situ remnants of the
population of planetesimals which accreted to form the Earth and other
terrestrial planets. The meteorites are samples of some of these asteroidal
bodies, but the specific source bodies of the individual meteorites or
meteorite groups have not been established. Thus the detailed meteoritic
evidence cannot be directly placed into a spatial context in the early
solar system. The meteorites provide a good "clock" for events and
conditions in the late nebular and early accretionary period of solar
system history, but they provide a poor "map".
The analysis of asteroidal visible and near-infrared (VNIR) reflectance
spectra compliments meteorite studies as a means of probing the late
nebular through early post-accretionary period of solar system history.
Although surface material characterizations from VNIR reflectance
spectrophotometry cannot achieve the detail or the level of sophistication
attainable in the laboratory analysis of meteoritic specimens, such remote
sensing studies do provide valuable and unique information on material
whose location in the early solar system can be established with some
certainty. Asteroid spectral studies also provide insight into the
incompleteness in the meteorite sample and into the types of additional
assemblages which are present on asteroids.
The particular type of material present on a specific asteroid provides
direct insight into the nebular or post-accretionary processes in the early
solar system at that particular semi-major axis. As the number of
characterized asteroids increases, so does the completeness of this "map"
of early processes and conditions. The combination of temporal constraints
from meteoritic studies and spatial constraints from asteroid spectral
characterizations provides the most powerful available means of
investigating the formation epoch.
This research program involves five complimentary efforts, including:
1) the development of quantitative interpretive calibrations and
procedures for the analysis of VNIR spectral data from laboratory
spectral studies of the meteoritic and meteorite-like assemblages which
are appropriate analogues to asteroidal materials [e.g. Gaffey, 1986a;
Cloutis et al^ 1986],
2) the re3uction and calibration of high precision VNIR telescopic data
of asteroids selected as particularly relevant to major issues in
asteroid or meteorite science,
3) the analysis of that asteroid VNIR spectral data using the most
sophisticated available interpretive calibrations and methodologies to
derive quantitative or semi-quantitative determinations of surface
mineralogy, phase abundance, and the mineralogic nature and lateral
extent of large scale lithologic units on the surface [e.g. Gaffey,
1983, 1984],
4) the utilization of these surface material characterizations to
constrain the evolution of individual asteroids, such as the presence
and intensity of any post-accretionary thermal events, the degree of
metamorphism or magmatic differentiation, and the nature and extent of
81
the subsequent collisional processes [e.g. Gaffey, 1984], and
5) the synthesis of this information on individual asteroids into models
of the inter-relationships between the various asteroid groups, the
post-accretionary collisional and thermal history of regions and
populations in the asteroid belt, and the implications for the accretion
and early evolution of the terrestrial planets [e.g. Gaffey, 1986b].
This multidiscipline approach, combining in a single program all the
major aspects of the work, has proven particularly efficient. It allows
each aspect to be better focused on the central issues. For example,
calibration work can be better directed toward satisfying the most pressing
needs in the interpretive area; the observations can be specific to the
questions being addressed rather than simply part an uncoordinated general
survey; and the analysis can effectively communicate its most urgent needs
for specific calibrations or for observations of a particular type.
However, in any asteroidal research beyond the most general type of
survey observations (i.e. in the first pass, one observes whatever happens
to be available), the quality of the scientific contribution which is
produced depends significantly on the appropriateness and sophistication of
the formulation of the original question. A major focus of this research
program is the issue of asteroidal thermal and collisional evolution.
The relative proportion of S-type asteroids decreases rapidly with
increasing semi-major axis across the asteroid belt. This has been
interpreted to indicate that either: a) the belt represents a transition
region between the zones of formation of ordinary chondrite-type and the
carbonaceous chondrite-type assemblages by nebular processes, or b) the
inner portion of the belt has undergone a strong post-accretionary heating
event which altered the original material in that region.
In the first alternative, the general population pattern in the belt is
the' fossil signature of the radial compositional variation of the accreting
solar nebula. In principle, once that pattern is understood, the
temperature and pressure of the nebula could be strongly constrained at
that particular distance from the proto-sun. In the second case, the radial
variation in the asteroid population is the signature of a radially
dependent heating event in the early post-accretionary period which was
superimposed upon the previous nebular material. It is plausible or
probable that such a heating event would have also affected the
planetesimals in the zones of the terrestrial planets. In that case, the
early thermal structure and processes (core formation, atmospheric
outgassing rates, etc.) of the terrestrial planets, which accreted from
such a population of planetesimals, would be a strong function of the
thermal state these planetesimals. An Earth accreted from warm,
magmatically differentiated planetesimals would have a profoundly different
early history than one accreted from C0I4, undifferentiated planetesimals,
even though both would have chondritie bulk compositions.
To date, essentially all available spectral evidence has strongly
indicated that the large S-type asteroids are predominantly thermally
evolved, magmatically differentiated bodies [see, Gaffey, 1984; Gaffey,
1986a, b). It is also now evident that the variation in UBV colors of the S-
type asteroids with semi-major axis (Dermott e;t al., 1985) is a
manifestation of a systematic mineralogic variation within the S-type
population, these objects becoming more metal- and pyroxene-rich and more
olivine-poor with increasing orbital distance (Gaffey, 1986b). This pattern
is the reverse of that expected from an undifferentiated, chondritie
planetesimal population. The relationships indicate that indeed the S-type
are primarily thermally-evolved bodies, and that the innermost S-type
objects underwent the most intense heating and differentiation while the
outer S-types underwent substantial reduction of oxidized iron with either
82
strong metamorphism or partial melting. Work is underway using 52-color
survey spectra XBell et al», 1985) to refine this picture.
Analysis of rotationaTiy resolved reflectance spectra has been used to
derive a litholoic unit map for 4 Vesta (Gaffey, 1983; and in preparation),
to investigate asteroid surface heterogeneity (Gaffey, et al., 1982), and
to test the chondritic affinities of 8 Flora (Gaffey, 1W4).
Present work is concentrated on 15 Eunomia, the largest S-type
asteroid. Observations in December, 1981 produced 101 24-filter spectra
(0,35-l.OOum) and 66 120-channel VNIR CVF spectra (0.65-2.55um), These
exhibit rotational variations which are being analyzed to determine the
nature of the hemispheric mineralogic variations which produce the
variations.
In addition, the flux variations in each filter or channel provides a
lightcurve for Eunomia at that bandpass. In a cooperative program with S.J.
Ostro (JPL), these monochromatic lightcurves are being inverted by his
technique (Ostro and Connelly, 1984) to derive an equivalent equatorial
profile of the body. Since the physical shape of the body is invariant,
differences between the profiles at different wavelengths must arise from
surface reflectance variations. Combined with the subhemispheric
mineralogic variations determined from the analysis of the spectra at
different rotational aspects, the variations of the complex profile within
and outside mineralogically diagnostic spectral intervals will permit
determination of lithlogic boundaries and, if Eunomia is an intact core
fragment, constrain the litho-stratigraphic structure within the parent
planetesimal. This should significantly increase our understanding of the
nature of the magmatic differentiation processes in asteroid sized bodies.
References
Bell, J,F, B.R, Hawke, P.D. Owensby, and M.J. Gaffey (1985) The 52-color
asteroid survey: Results and interpretation. Bull, A.A.S, 17, 729,
Cloutis, E,A,, M.J, Gaffey, T,L. Jackowski, K,L, Reed (1986) Calibrations
of phase abundance, composition, and particle size distribution for
olivine-orthopyroxene mixtures from reflectance spectra. J^ Geophys,
Res. 91. 11,641-11,653,
Dermott, S.F,, J, Gradie, and CD. Murray (1985) Variation of the UBV
colors of S-class asteroids with semimajor axis and diameter, Icarus 62 ,
289-297.
Gaffey, M,J. (1983) The asteroid (4) Vesta: Rotational spectral
variations, surface material heterogeneity, and implications for the
origin of the basaltic achondrites. Lunar Planet , Sci. XVI, 231-232.
Gaffey, M,J. (1984) Rotational spectral variations of Asteroid (8) Flora:
Implications for the nature of the S-type asteroids and for the parent
bodies of the ordinary chondrites. Icarus 60 83-114.
Gaffey, M.J, (1986a) The spectral and physicaT properties of metal in
meteorite assemblages! Implications for asteroid surface materials.
Icarus 66, 468-486,
Gaffey, fCT. (1986b) Evolution of the inner asteroid belt: Paradigms and
paradoxes from spectral studies, Meteoritics 21,
Gaffey, M.J., T. King, B.R. Hawke, and M.J. CintTla (1982) Spectral
variations on asteroidal surfaces: Implications for composition and
surface processes, in Workshop on Lunar Breccias and Soils and Their
Meteoritic Analogs (G,J. Taylor "ihrTXI Wilkenin^T'EaiX' TFT TirS2-02,
Lunar and Planetary Institute, Houston, pp. 40-43.
Ostro, S,J, and R. Connelly (1984) Complex profiles from asteroid
lightcurves . Icarus 57 , 443-463 .
83
SOUECE OF THE OPTICAL EED-SLOPE IN lEON-RICH METEOEITES, Daniel T. Britt,
Carle M. Pieters, and Peter H. Schultz, Department of Geological Sciences, Brown University, Box
1846, Providence, RI 02912.
The relationship between ordinary chondrites and the S-type asteroids is an unresolved issue
in meteorite science. S-type asteroids exhibit a positively red-sloped spectrum that has been
interpreted to indicate the presence of elemental iron on their surfaces. Some workers [1] suggest
that S-type asteroids represent undifferentiated parent bodies of ordinary chondrites on which
regolith processes have enhanced the spectral signature of metallic iron relative to laboratory
samples. Other workers [2] interpret the red-slope spectra -in S-types to indicate the presence of
relatively large amounts of elemental iron and suggest that the differentiated meteorites are better
analogs. The issue hinges on the amount and form of iron required to produce a spectral red-
slope.
Investigators using integrating-sphere spectrometers [3] note that iron-rich meteorites exhibit
a strong and characteristic red-slope in the visible and near-infrared wavelengths. Optical
constants measured at normal incidence for iron [4] and for iron-nickel alloys [5] also predict a
red-sloped spectrum. The effect of viewing geometry on the reflectance of iron meteorites has
been investigated using NASA's RELAB facility located at Brown University [6]. Figure 1 shows
the spectra of a clean, smooth portion of a nickel-iron meteorite taken from several viewing
geometries (these data are relative to the reflectance of halon and are displayed on a logarithmic
scale). A striking feature of these results is that iron's characteristic red-slope appears only in the
specular portion of the reflectance. All the non-specular geometries exhibit flat, featureless
spectra with the relative brightness being a function of the angular distance from the specular
reflection. These results suggest that the diffuse reflectance spectrum of iron and any iron-rich
complex surface is a weighted average of specular and non-specular components; the degree of
"redness" is a function of the amount of the specular component included in the measurement. As
a test of this hypothesis, bi-directional spectra were obtained of a complex iron surface (an 8 mm
crater produced in an experimental hs^pervelocity, metal on metal,, impact on the same iron
meteorite [7]). The spectrum of the crater on the meteorite exhibit a red-slope of 30%. The
proportions of shadow and specular, near-specular, and non-specular reflectance were estimated
for the meteorite crater, and a linear mixing model was applied using the spectra of figure 1. Less
than 0.2% of the specular component was required to accurately model the red-slope of the
meteorite crater.
CONCLUSIONS: The characteristic red-sloped spectrum of iron-rich meteorites is produced by
only the specular component of the reflectance. Complex metallic surfaces can be modeled as
linear mixtures of specular and non-specular components. It is the geometry of the metal on a
surface and its interaction with surrounding material, rather than the absolute amount of metal,
that determine the redness of resulting spectra. In order to distinguish between ordinary
chondrite smd differentiated parent bodies it is important to understand how regolith processes
affect the nature and form of metal on asteroid surfaces.
REFERENCES: [1] Feierberg, M.A., at al. ^
(1982) Astrophys. J. 257, p 361-372. [2]
Gaffey, M.J. and McCord, T.B. (1978) Space
Sci. Rev. 21, p 555-628. [3] Gaffey, M.J.
(1976) J. Geophys. Res. 81, p 905-920. [4]
Bolotin, G.A., et al. (1969) Phys. Metal.
Metallography 27, -p 31-41. [5] Sasovskaya,
LI. and Noskov, M.M. (1974) Phys. Metal.
Metallography 37, 'p 45-52. [8] Pieters, CM.
(1983) J. Geophys. Res. 88, p 9534-9544.
[7] Matsui, T. and Schultz, P.H, (1984) Proc.
Lunar Planet. Sci. Conf. 15th, in J. Geophys.
Res. 89, p C323-C328
FIGURE 1: Bi-directional refl^tance @f ircm
meteorite sample. Labels are m degrees
from specular reflection.
100-
10
■■-T"~ '«" 1 ""■
(
-
10
>«.«.iS
-
20
— —
40
-
60
_
.
•
1 i ......
f S
.
1
0.5 0.8 1.1 1.4 1.7
WAVELENGTH IN MICRONS
84
A SEARCH FOR SPECTRAL ALTERATION EFECTS IN CHONDRITIC GAS-RICH BRECCIAS «
J, F. Bell (Hawaii Institute of GeophysicSj Honolulu HI 96822) and Klaus
Keil (Institute of Meteoriticsj Univ. of New Mexicoj Albequerque NM 87131)
INTROKJCTION: The photons reflected by planetary surfaces do not
penetrate the surface by more than a few millimeterSj even in the most
transparent minerals. Thus the large number of visual and IR reflection
spectra of asteroids obtained with telescopes (and in the future by space-
craft) sample only the very uppermost regolith. However these spectra are
usually compared with lab spectra of solid meteorites derived from the
asteroidal bedrock. Solid meteorite fragments are pulverized to simulate
the scattering conditions in the upper regolith. This methodology impli-
citly assumes that the regolith-forming process on asteroids changes only
the particle size and does not introduce spectral effects due to glass for-
mations solar wind implantation^ etCe On the Moon^ very strong weathering
effects (mostly related to the formation of glassy agglutinates) exist in
the uppermost regolith which largely conceal the absorption features of
minerals, except in very fresh craters^ While it is generally agreed that
lunar-style glass formation is a very minor process in asteroidal regoliths
due to the lower impact velocities, several workers have proposed alternate
weathering effects to account for various mysteries connected with asteroid
spectral interpretation (e«g. Pieters 1984; T»V. V, King et. al » 1984;
McFadden and A'Hearn 1986; Weatherillj personal communications 1984). A
small subset of meteorites, the solar-gas-rich breccias, apparently
preserve portions of the uppermost regolith complete with alteration
effects characteristic of long exposure to the space environment. Typi-
cally, areas of fine-grained matrix in these meteorites are rich in
implanted solar wind gases, particle tracks left by galactic cosmic rays
and solar flares, and small fragments of exotic meteorites classes,
apparently projectile material which has survived low-velocity impacts.
These effects indicate that this matrix material once resided on the opti-
cal surface of an asteroid, was turned over by regolith gardening, lithi-
fied by an impact, ejected, and transported to Earth. (A similar history
is inferred for many lunar samples, including lunar meteorites). Large
clasts in the same breccias usually are devoid of these effects in their
interiors, which were protected from direct exposure to space. If any unk-
nown weathering effect exists in the uppermost regolith of asteroids which
significantly affects current interpretations of their composition, it
should be evident in a spectral difference between gas-rich matrix and
gas-poor clast interiors. We have carried out a systematic search for such
effects to test the hypothesis that they could make ordinary-chondrite
material resemble the common Class S asteroids.
MEASUREMENTS: Several samples of gas-rich breccias were selected,
including slabs of the Kapoeta howardite, the ordinary chondrites Dubrov-
nik, Cangas de Onis, and Dimmit. Numerous 0.8-2.5 micron reflection spec-
tra of selected areas on sawed or broken surfaces were measured with the
Planetary Geosciences Division spectrogoniometer. While these spectra are
not directly comparable to those of powdered samples (the continnua are
systematically bluer), comparisons within the data set should reveal any
spectral differences due to space weathering.
RESULTS: Kapoeta; Spectra of eucritic clasts resemble those of
powdered eucrites. Spectra of dark gas-rich matrix are very similar except
85
for slight reduction in band depths* Both lower albedo and reduced band
depths are probably due to mild shock effects. Dubrovnik: All spectra very
similar and resemble powdered ordinary chondriteSj with pyroxene band
depths reduced relative to eucrites due to presence of dispersed iron par-
ticles» Cangas de Onis: Spectra similar to Dubrovnik. A few highly
shocked dark veins have anomalously shallow bands ^ but gas-rich and gas-
poor regions are very similar, Diiranit: This meteorite contains a clast of
impact melt rock in a chondritic matrix, representing the most extreme form
of altered regolith. Spectra of this clast have very deep pyroxene bands*
They closely resemble those of the eucritic clasts in Kapoeta, and earlier
published eucrite spectra. This may be due to the lack of metal in the
Dimmit clast; it appears to be a fragment of a large melt pool in which the
metal component settled out»
IMPLICATIONS FOR ASTEROIDS: These results indicate that unknown
regolith processes do not conver the ordinary-chondrite parent bodies with
an altered layer exhibiting S-class spectral properties. This is con-
sistent with recent interpretations of the new Q-class of asteroids as the
ordinary-chondrite parent bodies. However j significant spectral effects do
occur in asteroid regoliths: A) Darkening and suppression of absorption
bands in highly shocked materials as seen previously in the so-called
"black chondrites"; B) Segregation of metal in large impact melt pools on
chondritic asteroids, which may have achondritic spectra. Neither of these
effects is likely to be significant in interpreting current integral-disk
spectra, but should be searched for in spectral maps returned by future
spacecraft.
REFERENCES s
Kingj T. ¥. V, (1984). Spectroscopic evidence of regolith maturation.
Meteoritics , 19 . 251-252.
McFadden, L. A., and M. F. A' He am (1986), Asteroid composition and
meteorite origins. In NASA TM-88383 ^ p. 14.
Pieters, C. (1984), Asteroid-meteorite connection: Regolith effects
implied by lunar reflectance spectra. Meteor itics . 19., 290-291.
86
ANALYTICAL ELECTRON MICROSCOPY OF FINE-GRAINED PHASES IN PRIMITIVE
INTERPLANETARY DUST PARTICLES AND CARBONACEOUS CHONDRITES.
Ian D. R. Mackinnon, Frans J. M. Rietmeijerj Department of Geology j University
of New Mexico s Albuquerque NM 87131 and David S. McKay, Mail Code SN4, NASA
Johnson Space Center, Houston TX 77058.
Exploration of the Solar System by spacecraft over the past few decades
has not only provided an enhanced appreciation of the geological diversity
within planetary bodies, but also a greater understanding of the origin and
evolution of our Solar System and perhaps others within the galaxy. A
fundamental component of this understanding involves the compilation and
analysis of chemical, physical and mineralogical data on the smaller bodies
within the Solar System. These bodies (or their fragments) probably contain a
record of processes occurring early in the evolution of the Solar System. In
general, the fine-grained fractions (i.e. <1 ym) of solid materials in
primitive meteorites and micrometeorites are the most likely hosts for this
record of early Solar System processes [1]. Any study of early Solar System
processes should involve (a) characterization of known extraterrestrial
materials intersecting Earth orbit [2,3], (b) an understanding of possible
contaminants both in low Earth orbit and the upper atmosphere [4,5], (c)
experimental verification of basic phenomena occurring in a stellar
environment [6] and (d) the production of realistic models on aspects of Solar
System evolution [1,7]. In the current phase of this research program, all
approaches have been utilized and highlights of this work are given below.
In order to describe the total mineralogical diversity within primitive
extraterrestrial materials, individual interplanetary dust particles (IDP's)
collected from the stratosphere as part of the JSC Cosmic Dust Curatorial
Program [8] have been analysed using a variety of AEM techniques.
Identification of over 250 individual grains within one chondritic porous (CP)
IDP shows that most phases could be formed by low temperature processes [2]
and that heating of the IDP during atmospheric entry is minimal and less than
600°C [2, 9]. This observation agrees with more recent entry temperature
estimates based upon solar flare track data. We have suggested that layer
silicates observed in these IDP's could have formed by low temperature
(cryogenic) alteration of precursor silicates [2]. In a review of the
mineralogy of IDP's [1], we also suggest that the occurrence of other
silicates such as enstatite whiskers is consistent with formation in an early
turbulent period of the solar nebula during which C/0 and Mg/Si ratios enhance
condensation at temperatures <1000°C.
Considerable attention has been paid to carbon-rich phases in IDP's and
at least three different forms have been identified [1]. Earlier work on
poorly-graphitised carbons (PGC's) has shown that the basal spacing of
synthetically-formed PGC varies with formation temperature [10]. This
structure-temperature relationship has been proposed as a new type of
cosmothermometer for primitive extraterrestrial materials [11]. This
temperature dependence has been confirmed for naturally occurring PGC's in
terrestrial environments [11,12], The lower graphitisation temperatures
predicted for carbonaceous chondrite and IDP PGC's also reflects the presence
of catalysts during the graphitisation process [11]. Use of this
cosmothermometer on observations of PGC's in carbonaceous chondrites or IDP's
[11,12] provides a record of the last temperature event experienced by the
87
host body. Other types of carbon observed in IDP's include the polymorph
carbon-2H, and we propose that this type of carbon formed by hydrous pyrolysis
of precursor hydrocarbons [14]. A collaborative laser microprobe study on
carbon-rich IDP'sj previously characterized by AEM, provides inconclusive
results on the presence of interstellar carbon [15] though they are consistent
with previous ion microprobe studies.
Micrometeorites captured in low Earth orbit have also allowed a
fortuitous calibration of orbital capture techniques for future dedicated
Shuttle or Space Station missions [16]. Components returned from the Solar
Maximum satellite have been exposed to space for about four years and thus,
contain a continuous record of the local micrometeorite and debris environment
by way of impact features (i.e. craters and penetration holes) with associated
projectile residues and adhered solids [3]. Electron optical analyses of the
returned Solar Max surfaces show that solid residue from the projectiles
survived impact dues in part to the multi-layer design of thermal blankets and
the variable approach velocities of Impacting bodies [17], Analysis of solid
particles associated with impact features readily allows the identification of
orbital debris^ such as paint particles [16] ^ as well as at least two types of
micrometeorites: Mg-Fe silicates and Fe-Ni sulphides [3,16] . Detailed AEM
observations on Mg-Fe silicate particles show that (a) they are Mg-rich
olivines with a chemical signature similar to olivine in CM chondrites and (b)
these olivines did not melt upon impact and survived impact without
appreciable shock metamorphic effects^ such as lattice distortion [18], These
observations provide an important database for the capture of pristine
micrometeorites by Earth-orbiting capture cells or comet coma dust sampling
devices [16s 17],
Experimental confirmation of fundamental chemical and physical processes
in a stellar environments such as vapor phase condensations nucleationj and
growth by annealing j is an important aspect of astrophysical models for the
evolution of the Solar System, For examples the microstructural development
of refractory smokes can provide significant constraints on the kinetics of
particle growth and accumulation in a stellar environment. Characterization
of laboratory-produced smokes has shown that both infra-red spectroscopy and
X-ray diffraction fail to detect the very initial stages of crystallite
development during high temperature annealing of "amorphous" smokes [6].
Detailed AEM study indicates that microcrystallites of forsterite may directly
condense from an MgO-SiO vapor phase system or form metastably shortly after
condensation [6]. With annealing, both compositional and structural
transformations of the MgO-SiO smokes occur and have been documented using the
AEM [6], Similar textural and structural observations have been presented for
the ultra fine-grained minerals in four anhydrous chondritic IDP'sj and are
interpreted as evidence for annealing in the early history of the Solar System
[15sl9s20s25].
An inherent limitation of terrestrial-based laboratory experiments on
particle condensation, nucleation and growth from a vapor phase is the
influence of the Earth's gravity during smoke production. To overcome this
experimental limitations suggestions on the development of a particle (or
"dust") facility on board the Shuttles and ultimatelyj upon the Space Station
have been proposed [21s22]. This potentially rich area for experimental
confirmation of fundamental astrophysical concepts has also received attention
with respect to understanding solar nebula physico-chemical processes such as
turbulence and particle aggregation [23524].
88
On a larger scales the possible relationships between chondritic IDP's
and chondrite meteorite components have also been investigated [ls25]. In one
case, there may be mineralogical similarities between an IDP and the matrices
of CO/CV chondrites or unmetamorphosed unequilib rated ordinary chondrites
[25]. Howevers in general, a detailed comparison of chondritic IDP and
carbonaceous chondrite mineralogies shows significant differences between the
types of silicate minerals as well as the predominant oxides [1]. In a
continuing effort to characterize the fine-grained matrix of carbonaceous
chondrites, detailed structural and morphological studies of the mineral
tochilinite have been presented [26]. This work arises from the suggestion
that the commonly-known "poorly-characterized-phases" (PCP's) can in fact be
identified as members of the mixed-layer tochilinite mineral group or as
coherent intergrowths of tochilinite and serpentine [27]. This is a critical
observation for models of CM matrix formation as tochilinite paragenesis,
though incompletely studied, appears restricted to very specific environments
on Earth [28].
References s
1. Mackinnon IDR and Rietmeijer FJM (1986) Rev, of Geophysics , in press.
2. Rietmeijer FJM and Mackinnon IDR (1985) J. Geophys. Res ., 90, D149-D155.
3. Rietmeijer FJM et al., (1986) Adv. Space Res . , in press.
4. Mackinnon IDR (1986) LPI Tech Rpt. 86-05 , 68-72.
5. Rietmeijer FJM (1986) LPI Tech. Rpt . 86-05, 80-82.
6. Rietmeijer FJM et al. , (1986) Icarus , 65, 211-222.
7. Rietmeijer FJM (1986) Nature , 313, 293-249,
8. Mackinnon IDR et al., (1982) J. Geophys. Res. , 87, A413-A421.
9. Mackinnon IDR and Rietmeijer FJM (1984) Nature, 311, 135-138.
10. Rietmeijer FJM and Mackinnon IDR (1985) Lunar & Planet. Sci. XVI , 700-701.
11. Rietmeijer FJM and Mackinnon IDR (1985) Nature, 315, 733-736.
12. Lumpkin GR (1986) Lunar & Planet. Sci. XVII , 504-505.
13. Lumpkin GR (1986) Lunar & Planet. Sci. XVII , 502-503.
14. Rietmeijer FJM and Mackinnon IDR (1986) Nature , in press.
15. Carr, RH et al,, (1986) Meteoritics , 21, in press.
16. McKay, DS et al., (1986) LPI Tech Rpt. 86-05 , 72-75,
17. Rietmeijer FJM (1986) Rpt on Multi-Comet Mission , in press.
18. Blandford GE et al. , (1986) Lunar & Planet. Sci. XVII , 56-57.
19. Rietmeijer FJM and McKay DS (1986) Lunar & Planet. Sci. XVII , 710-711.
20. Rietmeijer FJM (1986) Meteoritics , 21, in press.
21. Nuth JA et al., (1985) SSPEX Workshop , 62-63,
22. Nuth JA et al., (1985) In: Micro-gravity Particle Research on the Space
Station , 2-8.
23. Mackinnon IDR and Rietmeijer FJM (1985) In: Rpt. on a Workshop on
Experimental Cosmo-chemistry in the Space Station , 15-16.
24. Iversen J et al., (1985) SSPEX Workshop , 7-12.
25. Rietmeijer FJM and McKay DS (1985) Meteoritics , 20, 743-744,
26. Zolensky ME and Mackinnon IDR (1986) Am. Mineral ., 71, 1201-1209.
27. Mackinnon IDR and Zolensky ME (1984) Nature , 309, 240-242.
28. Zolensky ME (1984) Meteoritics, 19, 346-347.
ASTEROID LIGHTCURVE INVERSION
Steven J. Ostro, Jet Propulsion Laboratory, Pasadena, CA 91109
Robert Connelly, Cornell University, Ithaca, NY 14853
One of the most fundamental physical properties of any asteroid is its
shape, i.e., its dimensions. Lightcurves provide the only source of shape
information for most asteroids. Unfortunately, the functional form of a
lightcurve is determined by the viewing/illumination geometry ("VIG") and
the asteroid's light- scattering characteristics as well as its shape, and
in general it is impossible to determine an asteroid's shape from
lightcurves (1) . The best one can do is to derive a shape constraint that
is useful and takes advantage of all the information in the lightcurve.
We have introduced a technique called convex-profile inversion (CPI)
that obtains a convex profile, P, from any lightcurve (2). If certain
ideal conditions are satisfied, then P is an estimator for the asteroid's
"mean cross section", C, a convex set defined as the average of all cross
sections C(z) cut by planes a distance z above the asteroid's equatorial
plane. C is therefore a 2-D average of the asteroid's 3-D shape. The
ideal conditions are that (A) all C(z) are convex; (b) the asteroid's
scattering law is geometric, so brightness is proportional to the
projected visible, illuminated area; (C) the VIG is equatorial, i.e., the
asteroid- centered declinations 5g and 5c of the Earth and Sun are zero;
(D) the solar phase angle ^ 9^ .
The first three conditions are unlikely to ever be satisfied exactly,
but the issue here is the extent to which their violation degrades the
validity of P as an estimator for C. Laboratory simulations suggest that
modest, "topographic" concavities play a relatively minor role in
determining the form of a lightcurve (3) . Similarly, numerical
experiments indicate that systematic errors introduced by small ( several -
degree) violation of Condition C are not severe (4) . The bulk of
available polarimetric and spectrophotometric data show that whereas
"hemispheric" albedo variations can be detected at about the several
percent level for several asteroids [e.g., (5)], the forms of most
broadband optical lightcurves seem less sensitive to surface heterogeneity
than to gross asteroidal shape.
Geometric scattering is expected to be an excellent approximation
close to opposition (6) , but a poor approximation far from opposition (7) .
This is an unfortunate circumstance, because CPI's ability to reveal odd
harmonics in G improves as <f) increases. The systematic error introduced
by non- geometric scattering will depend on the 3 -dimensional shape as well
as on the VIG. Hence, the nature and magnitude of this error will be to
invert several lightcurves obtained under nearly ideal VIG but at a
variety of solar phase angles , use the weighted mean profile as an
estimate of C, and use the variance in the profiles to gauge hte severity
of systematic error.
What can an opposition lightcurve tell us about an asteroid's shape?
At <f> = 0, CPI yields a profile Pg as an estimator for Cg , the
"symmetrization" and the ratio /3, of the profile's maximtim breadth to its
minimiim breadth, remain intact.) Moreover, Condition A need not hold at
opposition, and if it does not, then Cg is the sjrmmetrization of the
asteroid's mean convex envelope, or "hull".
90
Opportunities for reliable estimation of Cg should be much more
abundant than those for estimation for C for several reasons. First, the
VIG required for reliable estimation of Cg occurs much more frequently
than that for estimation of C. Second, to assess how close the VIG is to
ideal, we want to know the direction of the asteroid's spin vector when
estimating C, but we just need to know the direction of the asteroid's
line of equinoxes when estimating Cg . Third, as noted above, geometric
scattering is most valid close to opposition. Finally, as shown by
Russell (1) , it is easy to test the hypothesis that the ideal conditions
required for reliable estimation ofCg (6g=5g=^=0 and geometric
scattering) actually hold; if the lightcurve has any odd harmonics, the
conditions are violated.
Cg is the sjomnetrized average of all C(z) and constitutes the optical
extraction of shape information from an opposition lightcurve, just as C
constitutes the optimal extraction of shape information from a non-
opposition lightcurve. If an estimate of C were free of systematic
errors, its s3nmnetrization would look the same as an estimate of Cq , so we
can use opposition lightcurves to qualify the interpretation of non-
opposition lightcurves.
This profile is our estimate of Cg for
asteroid 624 Hektor from CPI of a
lightcurve obtained by Dunlap and
Gehrels (8) at ^ - 4°. The pole
directions estimated by those authors
and by Magnusson (9) indicate that |5g|,
|5g| < 10°. The constancy of Hektor 's
color indices with rotational phase (8)
results of Russell's Fourier test, and
the goodness of fit of CPI's model
lightcurve to the data concur in
supporting the expected high reliability
of this estimate.
The profile has /3 = 2 . 5 and is distinctly non- elliptical. Since the
mean cross section of an ellipsoid rotating about a principal axis is an
ellipse, our results suggest that neither the asteroid nor its convex hull
are ellipsoids. On the other hand, our results are quite consistent with
many other models for Hektor' s 3-D shape, including a cylinder with
rounded ends (8), a dumbbell (10), and various binary configurations.
REFERENCES : 1) Russell H.N. (1906), Astrop . J. 24, 1-18. 2) Ostro S. J.
and Connelly R. (1984), Icarus 57, 443-463. 3) Barucci M.A. and
Fulchignoni M. (1981), in Asteroids , Comets , Meteors (C.-I. Lagerkcist and
H. Rickman, Eds.), pp. 101-105. 4) Ostro S. J., Dorogi M.D. , and Connelly
R. (1985), Lunar Plan . Scl. 16 Abstracts , 637-638. 5) Gaffey M. J.
(1984), Icarus 60, 83-114. 6) French L. M. and Veverka, J. (1983),
Icarus 54, 38-47. 7) Uxame K. and Bowell E. (1981), Astron. J. 86, 1694-
1704. 8) Dunlap J. L. and Gehrels T. (1969), Astron. J. 74, 796-803, Fig.
8. 9) Magnusson P. (1986), Icarus , in press. 10) Hartmann W. K. and
Cruikshank D. (1978), Icarus 36, 353-366. 11) Weidenschilling S. L.
(1980), Icarus 44, 807-809.
91
Accumulation of the Planets
G. W. Wetherill, DTM, Carnegie Institution of Washington
A. Early Stages of Planetary Growth.
In modelling the accumulation of planetesimals into planets, it is appropriate
to distinguish between two stages:
(A) An early stage, during which ~ 10 km diameter planetesimals accumulate
locally to form bodies ~ 10^^ g in mass. During this stage, it is useful to treat the
bodies as particles, analogous to gas molecules in the kinetic theory of gases.
(B) A later stage in which the ~ 10^^ g planetesimals accumulate into the final
planets. During this stage it is likely that bodies will become well-mixed as a conse-
quence of radial excursions and the "giant" impacts will occur. During this stage it
is better to trace the orbital evolution of the larger individual bodies.
Previous work on the early stage has been extended by use of new expressions
developed by G. R. Stewart of the University of Virginia to describe the changes in
velocity of the bodies as a consequence of mutual gravitational perturbations, col-
lisions, and gas drag. The contribution of one of these terms, "dynamical friction",
has not been included in any earlier calculations, and all previous workers have ne-
glected at least one of the other terras as well.
In the terrestrial planet region, an initial planetesimal swarm corresponding to
the critical mass of dust layer gravitational instabilities is considered. It is assumed
these bodies have an initial exponential mass distribution dN a e""*/*"" dm where
nxo = 3 X 10-^^ (11.5 km diameter) and an initial relative velocity of 11 m/sec, and
the gas density to be 1.18 x 10~^ g/cw? . The continuous distribution is modelled
as an assemblage of 10^ bodies divided into 22 "batches", each batch containing
bodies of equal mass distributed over a'zone of width Aa = 0.02 A.U. The size and
velocity evolution of each batch is followed through successive time steps by calcu-
lating the effects of gravitational perturbations, collisions, gas drag, merger, and
fragmentation. When appropriate, the effect of low velocities on the gravitational
cross- section was included.
For a continuous initial size distribution, as assumed by previous workers, an
orderly growth is found, the early stage of accumulation ending when most of the
mass of the terrestrial planet region is concentrated in ~3000 bodies ranging from
~ 10^* g to ~ lO'^^ grams in mass at the time (~ 10^ years) that eccentricities rose
to ~.01.
"Runaway accretion" , in which most of the mass of the zone becomes concen-
trated in a single low eccentricity > 10^^ g body in < 10^ years, was found to re-
quire a discontinuous initial mass distribution, i.e. a "seed" at the upper end of
the distribution. The size of the seed required is somewhat dependent on the pa-
rameters assumed for gas drag and fragmentation. For the more plausible values, a
"seed" 2 to 3 times the mass of the 2nd largest body is required. The presence or
absence of such seeds is not at present predictable. For this reason, it is permissi-
ble to speculatively entertain models in which volatile-rich seeds are formed beyond
the "snow line" of the solar nebula at ~5 A.U., but not in the asteroid belt and the
terrestrial planet region. This could be a factor in facilitating the rapid growth of
planetesimals of Jupiter and Saturn, despite the relatively long accumulation times
conventionally associated with these large heliocentric distances.
B. Late Stages of Planetary Growth,
Our previous work on this problem has concentrated on the larger terrestrial
planets. Earth and Venus. Because of their smaller size and less "averaging", the
92
present state and chemical composition of Mercury and Mars may be expected to
be more diagnostic of events during planet formation.
In order to understand better the accumulation history of Mercury-size bod-
ies, 19 new Monte-Carlo simulations of terrestrial planet growth have been calcu-
lated. Three cases are presented, involving different assumptions regarding the ini-
tial state of the final stage of planetary accumulation and the degree and ease with
which plaxiets can be collisionally disrupted. It is found that the same conditions
that lead to Mars-size giant impacts on Earth and Venus imply a more catastrophic
fragmentation history for Mercury-size bodies and the fragments from which they
accumulated. In accordance with these results, it may be speculated that the large
iron core of Mercury may be a consequence of a giant impact that removed the sil-
icate mantle frora a previously differentiated body. Less extreme fragmentation
events may also contribute to chemical fractionation processes for which bodies in
this mass range may be especially susceptible. It is also found that planets of the
size and position of Mercury will accumulate material originating over the entire
terrestrial planet range of heliocentric distances. This tends to reduce the relative
importance of more primordial chemical fractionation processes.
C. Primitive Earth-approaching Bodies (Meteorites, Apollo-Amors, Me-
teors).
A Monte Carlo technique has been used to investigate the orbital evolution of
asteroidal collision debris produced interior to 2.6 A.U. It is found that there are
two regions primarily responsible for production of Earth-crossing meteoritic ma-
terial and Apollo objects. The region adjacent to the 3:1 Jovian commensurability
resonance (2.5 A.U.) is unique in providing material in the required quantity and
orbital distribution of the ordinary chondrites. This region should also supply a
comparable preatmospheric flux of carbonaceous meteorites. This work has been
extended to include the innermost asteroid belt. The innermost asteroid belt (2.17
to 2.25 A.U.), via the vq secular resonance, provides a flux ~9% that of the ordi-
nary chondrites, and appears to be the strongest candidate for the basaltic achon-
drite source region. It is unlikely that a signiflcant number of meteorites originate
beyond 2.6 A.U. It is speculated that enstatite achondrites are derived from the
Hungaria region, interior to the main belt, and that iron and stony- iron meteorites
originate from many main belt sources interior to 2.6 A.U.
The same techniques have been extended to include the origin of Earth-
approaching asteroidal bodies. It is found that these same two resonant mechanisms
predict a steady-state number of Apollo- Amor about 1/2 that estimated based on
astronomical observations. There axe some types of observed Apollo- Amor orbits
that are observationally overpopulated by a factor of 10 - 30 when compared with
predictions of the asteroidal source model. These deficiencies are remedied by pos-
tulation of additional production of Apollo- Amor objects as outgassed comets at a
rate of ~ 10~^ yr~-^.
Particularly puzzling questions are raised by sun-approaching bodies, and at-
tention is being given to these. Two Apollo objects have perihelia <0.2 A.U. and
aphelia <2.5 A.U. (1566 Icarus and 1983TB). The latter is associated with the
Geminids, the second most prolific of the annual meteor showers. At least 1% of the
Prairie Network fireballs are in similarly small orbits; because of selective effects the
fraction must be higher (e.g. Geminids were excluded in reducing the data). These
orbits are unstable on a time scale of ~ 10^ years and a fairly productive source is
required. None of the known mechanisms (either cometary or asteroidal) for supply-
ing Earth-crossing material can accomplish this.
93
This lack of correspondence with known cometary and asteroidal material has
been extended by examination of the atmospheric trajectories of six -5 to -8 magni-
tude meteors: three bright Geminids reported by Jacchia and three non-Geminid
Prairie Network fireballs in similar Sun- approaching orbits. These objects, both
Geminid and non-Geminid, also show physical similarities to one another. All of
them are Ceplecha and McCrosky type I fireballs, in this way we have identified as
ordinary chondrites. Their ratio of photometric/dynamic mass is at least as small,
and often smaller than ordinary chondritic fireballs, a further indication of strong,
dense bodies. They are clearly unlike typical Taurids (Comet Encke) and other
cometary fireballs which fragment more easily and are probably of lower density.
These bodies, however, have much lower ablation coefficients (<7 ~ 5 x 10"" -^^
sec^ /cnr^) than ordinary chondritic asteroidal fireballs (cr ~ 2 x lO"'^'^ sec'^/cm'^).
If loss by ablation were all that mattered, a Geminid would retain 4% of its ini-
tial mass while penetrating the atmosphere, despite its entry velocity of 36 km/sec.
Nevertheless, they fail to fully decelerate in the atmosphere, instead they disappear
at velocities of 17-32 km/sec at dynamic pressures of ~ 10^ dynes/cm.'^. For this
reason, their breaking strength appears to be no greater than ordinary chondrites,
despite all this other quantitative evidence for "toughness". The relatively small
mass (~ 1 — lOOg) of these bodies makes it difiicult to know whether or not any pos-
sibly collectible material survives their apparent disappearance. Measurement and
reduction of larger Geminids on fireball network photographs could permit address-
ing this question.
94
CHAPTER 4
COSMOGONY AND DYNAMICS
Protostellar disks and the Primitive Solar Nebula
P, M, Cassen, J. B, Pollack, T, Bunch
0. Hubickyj , P. Moins
NASA Ames Research Center
Moffett Field, CA 94035
C. Yuan
City College of New York
New York, NY 10031
(1) Turbulence in the solar nebula . This is a new project that we
expect to figure prominently in our future work. The objective is to
obtain quantitave information on the turbulent transport of mass,
angular momentum and energy under the conditions that characterized the
solar nebula, by direct numerical calculations. These calculations have
been made possible by research conducted on new supercomputers (Cray XMP
and Cray 2) by the Ames Computational Fluid Dynamics Branch, with whom
we have established a strong collaboration. Techniques have been
developed that permit the accurate representation of turbulent flows
over the full range of important eddy sizes ~ from the largest scales
at which energy is deposited in turbulent motions, down through the
inertial sub-range, to the small scales at which viscous dissiptaion
becomes important. So far, these techniques have been applied (and
verified) primarily in mundane laboratory situations, but they have a
strong potential for astrophysical applications.
Most current models of the solar nebula are based on the hypothesis that
turbulent stresses drive its evolution. But disagreement exists on the
effectiveness, or intensity, of the turbulence. All models are burdened
by untested assumptions regarding the details of the driving mechanisms,
the vertical distribution of dissipated energy, and the all-important
effects of rotation. Our approach is to conduct a sequence of numerical
experiments to evaluate the Reynold's stress tensor, turbulent heat
transfer rate, turbulent dissipation rate, and turbulent kinetic energy
spectrum, as functions of position, for conditions relevant to the solar
nebula. Emphasis is placed on the variation of these properties with
appropriate nondimensional quantities (such as the Rayleigh, Taylor, and
Prandtl numbers), so that relations can be derived that will be useful
for disk modeling under a variety^ of hypotheses and initial conditions.
In these experiments, we intend to examine separately the
characteristics of three potential sources of turbulence in the
nebula: thermal convection, mass infall from outside the nebula, and
shear in the angular velocity.
The codes that were developed for engineering purposes require
modification and generalization in a number of areas to realize their
full potential for nebula studies, and this is where our efforts have
been devoted so far. Calculations of Benard convection and convection
driven by internal heat sources have been performed. Diagnostics have
been devloped that permit easy examination of turbulent correlation
coefficients, energy spectra, mean variable profiles, all evaluated at
selected locations in the computational grid. The convection results
97
have been compared with detailed experimental results and those of other
numerical computations to the point where we are confident that our
resolutions, computational grid size, integration time, etc. are
adequated. We are now embarking on a series of experiments that will
become progressively more general and complex, as we (1) obtain results
for parameter ranges more relevant to nebula studies (e.g, low Prandtl
numbers) that have not been previously examined, and (2) Include more
physical effects that will be necessary for our goal of understanding
nebula turbulence.
(2) Self-gravitating Disk Models . Most models of the solar nebula ignore
the self-gravity of the disk compared to that of the protosun. In many
circumstances this is a good approximation, but it is clearly inadequate
for gravitational instability of the disk, to form spiral density waves,
planets or multiple star systems, say, should treat the gravitational
field of the disk self-consistently. So far, this has only been
attempted in studies that use hydrodynamical codes to calculate the
collapse of protostellar clouds. However, the results of theses
calculations are usually not the formation of a solar nebula-like disk,
but rings or ring fragments (see review by Bodenheimer and Black, in
Protostars and Planets , U. of Arizona Press, 1978), or in some cases,
thick disks that are subject to non-axisymraetric instabilities (Boss,
Icarus , 61 , 3, 1985). In some cases, an ad hoc turbulence is added to
transport angular momentum outward, allowing a disk to form (Morfill et
al., in Protostars and Planets II , U. of Arizona Press, 1985). These
results contrast with semi-analytic calculations of cloud collapse by
Terebey et al, ( Ap. J. , 286 ,529, 1984) , which suggest that a slowly
rotating cloud can collapse directly to a disk.
In order to examine the ring/disk formation question, as well as to
provide a convenient method for studying self-gravitating disk models for
other purposes, we have devised a method whereby (nearly) arbitrary
distributions of mass and angular momentum in spherically symmetric
protostellar clouds can be mapped into corresponding distributions in
infinitely thin self-gravitating disks, under the assumption of strict
angular momentum conservation. The method is based on Toomre's ( Ap. J. ,
138 , 385, 1963) Bessel integral formulation for the mass distribution in
flattened galaxies. Solution of an integral equation, by an iterative
methods is required; but what appears to be an excellent first
approximation for the mass distribution in the disk can be found
analytically for many cases. These preliminary results agree very
closely with the corresponding results of Cassen and Moosman ( Icarus , 48 ,
353* 1981) for inviscid disks. Furthermore, they suggest that disks
(rather than rings) can form a wide range of initial cloud conditions,
even at high rotation rates, as long as enough accretional energy can be
radiated away so that the thin disk approximation is valid. Calculations
of the cooling of the post-accretion shock gas indicate that this is the
case. The implication is that numerical collapse calculations must
resolve the accretion shock cooling region very well in order to get
accurate results, even on the basic question of ring- vs. disk formation.
98
(3) Analysis and interpretation of meteoritic Inclusions . Examination of
the components of the Allende meteorite has continued, along with the
development of a model postulated to explain the observed thermal
processing of CAIs (see Bunch et al.. Lunar Planet. Scl. XVI , 97, 1985;
Cassen et al. Ibid. 117, 1985; Bunch et al., Lunar Planet. Scl. XVII , 87,
1986). The essence of the hypothesis is that the secondary processing
exhibited by these objects resulted from an episode of aerodynamic
heating most plausibly attributed to entry into the atmosphere of a large
planetislmal. Mlcroprobe and SEM analyses performed this year indicate
that "classic" chondrules and f erro-magnesium aggragataes in Allende may
also have experienced thermal processing related to the growth processes
of a parent body. Details are provided in the attached Appendix,
Synthesis of this data, along with an evaluation of other heating
mechanisms, are currently in progress. The implications of our
conclusions with regard to the Allende parent body are that It was a
large (~1,000 km radius) body possessing a substantial (but perhaps
transient), dusty atmosphere.
99
Timescales for Planetary Accretion and the
Structure of the Protoplanetary Dxsk
Jack J. Lissauer (U, California, Santa Barbara)
Models of planetary accretion which assume the
mass of condensable matter in the protoplanetary
disk was equal to that present in the planets
today predict accretion timescales for the outer
planets r^^ 10 years. Such timescales are in-
consistant with observations of star forming
regions^ which suggest that most of the gas in
disks around one solar mass stars is removed in
a few X 10 years. This paper outlines a uni-
fied scenario for solar system formation con-
sistent with astrophysical constraints, Jupiter's
core could have grown by runaway accretion of
planetesimals to a mass sufficient to initiate
rapid accretion of gas in times of order of
5 X 10 - 10 years, provided the surface density
of solids in its accretion zone was at least 5-10
times greater than that required by minimum mass
models of the protoplanetary disk. The inner
planets and the asteroids can be accounted for in
this picture if the surface density of the solar
nebula was relatively uniform (decreasing no more
rapidly than r~^'3) out to Jupiter's orbit. The
total mass of the protoplanetary disk could have
been less than one tenth of a solar mass pro-
vided the surface density dropped off more steep-
ly than r~^ beyond the orbit of Saturn, The
outer regions of the nebula would still have con-
tained enough solid matter to explain the growth
of Uranus and Neptune in 5 x 10^ - 10*^ years,
together with the coincident ejection of comets
to the Oort cloud. The formation of such a
protoplanetary disk requires significant trans-
port of mass and angular momentum, and is con-
sistent with viscous accretion disk models of
the solar nebula.
100
Formation of Giant Molecular Clouds in Global Spiral Structures:
The Role of Orbital Dynamics and Cloud-Cloud Collisions.'
W. W. Roberts. Jr., G. R. Stewart (U. Virginia)
We investigate the different roles played by orbital dynamics and
(^issipative cloud-cloud collisions in the formation of giant molecular
clouds (GMCs) in global spiral structures. The interstellar medium
(ISM) is simulated by a system of particles, representing clouds,
which orbit in a spiral-perturbed, galactic gravitational field.
Detailed comparisons are made between the results of cloud-particle
simulations in which the cloud-particles collide inelastically with
one another and give birth to and subsequently interact with young
star associations and the results of stripped-down simulations in
which cloud-cloud collisions and star formation processes are omit-
ted. Large ^GMC-like" associations of smaller clouds are efficiently
assembled in spiral arms and subsequently dispersed in interarm
regions largely by the orbital dynamics alone. The overall magni-
tude and width of the global cloud density distribution in spiral
arms is very similar in the collisional and collisionless simulations.
The' results suggest that the assumed number density and size dis-
tributions of clouds and the details of individual cloud-cloud colli-
sions have relatively little effect on these features. In the simula-
tions with shorter mean free paths, pronounced shock-like density
and velocity profiles occur. In the simulations with longer mean
free paths and in the collisionless simulations, we find more sym-
metric, less shock-like density and velocity profiles. The natural
tendency of orbits to crowd together in spiral armis is enhanced by
the temporary trapping of clouds in spiral arm potential minima
for periods up to 50 Myr. Dissipative cloud-cloud collisions play
an important steadying role for the cloud system's global spiral
structure. Dissipative cloud-cloud collisions also damp the relative
velocity dispersion of clouds in massive a^oclations and thereby aid
in the effective assembling of GMC-like complexes. The assembly
of these GMC complexes from smaller clouds is remarkably efficient
even if collisional coalescence of individual clouds is inefficient.
'Work supported in part under NSF grant AST-82-04256 and
NASA grant NAGW-929-,
101
ANALYSIS OF PLANETARY EVOLUTION WITH EMPHASIS ON
DIFFERENTIATION AND DYNAMICS
William M. Ka^la and William L Newmaa
Uffliversity of Califomia. Los Angeles
Task #1: Collapse of tie Proto Solar System Cloud
In order to address the early stages of nebula evolution, we are developing a three-
dimensional collapse code which includes not only hydrodynamics and radiative transfer,
but also the effects of ionization and, possibly, magnetic fields. We intend to focus on
understanding the properties of a protosteliar cloud which lead to the formation of a Jupiter-
like second largest body, since such a body would have had a significant influence on the
subsequent dynamics of the circumsolar material from which the planets formed.
Ultimately, we intend to couple this collapse code with the accretionary N-body being
developed in task #2 which addresses the problem of planet formation.
Over the past few months, we have developed a numerical model for protosteliar
collapse similar to one developed by Boss in 1979. We are developing a hydrodynamic
scheme that minimizes the effect of numerically-induced artificial diffusion and expect to
improve upon the overall accuracy by replacing the first order finite difference scheme with
a higher order difference scheme or, preferably, a spectral or Galerkin technique.
In order to define the specific numerical technique that we intend to use, we have
surveyed the current literature for a method which minimizes boundary errors while
remaining computationally efficient. Glatzmaier (1984) developed spectral and Galerkin
techniques to model magnetohydrodynamic problem in solar physics, methods that are
well-suited to the protosteliar collapsfe problem. However his model accommodates only
the shell of a sphere. Currently, we are adapting this technique to full spherical geometry
by reducing the lower boundary to the sphere's center. Thus we hope to minimize the
problem of artificial diffusion and develop an accurate means for dealing with the inner
region of the nebular cloud where the important physical changes take place.
Task #2: Formation, of the Post-Jovian Planets
As part of our examination of solar system evolution, we have developed an N-
body code describing the latter stages of planet formation from the, accretion of
planetesimals. It is based on Aarseth's (1972) scheme and is a sixth order, three
dimensional code which solves the equations of motion precisely for N mutually attracting
bodies. In practice, the run-time for such a code varies as N'^. It is our goal to reduce this
cubic exponent through efficient vectorization.
To test our code for accuracy and run-time efficiency, and to develop a stronger
theoretical foundation for our study, we have used it to study problems in orbital dynamics.
This summer, we examined resonant interactions between Neptune and Pluto, with Jupiter
as the single other perturbing body. We observed significant exchange of angular
momentum over the equivalent of 1.9 million years. Our results support conclusions made
by others (Nacozy and Diehl; 1978, Williams and Benson, 1971), who found that
resonance improved the stability of the Neptune-Pluto system. We have reproduced
periodicities they found in eccentricity and inclination. Understanding such resonant
af'Fraf-'K? ■« p Vs o c-« j^ t ir\ n sn^^it-ot"*! M ri-< •^ /T *vi a ^^ tn o •^ ■* c •*! o ^rrWur^V*
*v» «rST7
In the beginning of this year, we converted our code to run on the CRAY
computers at Los Alamos National Laboratory, and later at the San Diego Supercomputer
Center. The Neptune-Pluto problem took nearly two hours of computer time. In
anticipation of problems with many more bodies, we have updated the subroutines which
perform the force calculations to enhance vectorization. These routines occupy
approximately 90% of the CPU time used by our program. Test runs show that the
combined effect of moving the code onto the CRAY and vectorizing the force routines has
been a twelvefold increase in speed over that of scalar machines.
The next step in suiting our code for planetesimal accretion is to invoke the ability to
handle collisions. We have now written the routines to identify and service two-body
interactions. With this capability, we intend next to compare our results to those of other
investigators, such as Wetherill (1985). WetheriLtl simulated the accumulation of 500
moon-sized bodies in a gas free environment. His Monte Carlo simulations typically
produced four roughly earth sized planets in the order according to mass like that of our
own solar system. We shall examine the sensitivity of the outcome (i.e. number and
location of planets) to initial conditions. We also intend to develop more precise values for
certain parameters, such as gravitational cross section, collision frequency and impact
energies, which are essential to analytical treatment of the problem.
Task #3: Regional Tectonics of Venus
We are performing a regional analysis of the correlation between the gravity and
topography fields of Venus, in an effort to determine the small and intermediate scale
subsurface structure. This subsurface structure will provide us with valuable information
concerning the depth of compensation, flexural rigidity of the lithosphere, thermal profile,
and ultimately, the tectonic regimes characterizing Venus. Knowledge of Venus tectonics
will then allow us to make inferences concerning reasons for the apparent differences
between the earth and Venus in terms of global tectonics.
The analysis of the gravity and topography of Venus is done using a two-
dimensional Fourier admittance function technique, developed by Dorman and Lewis
(1970) and McNutt (1979). This technique allows us to analyze the correlation between
gravity and topography at different wavelengths, and to compare these correlations with
model admittances. The main complication involved with this technique is the nature of the
Venus gravity measurements. While a vertical gravity field is required for this admittance
analysis, the Pioneer Venus Orbiter data consists of accelerations measured in the line-of-
sight (LOS) direction - that is, accelerations along a line between the Venus surface, the
spacecraft, and earth.
We have written the computer programs necessary to convert this LOS gravity field
to vertical gravity using quadratic sum minimization, an inverse method described by Kaula
(1966), Jackson (1979), and Jackson and Matsu'ura (1985). This method combines the
measured data, using a matrix representing the complete LOS observation geometry, the
LOS data, and a covariance matrix of the measurement errors, with a priori information, in
the form of a covariance matrix of resolving errors. This method involves two major parts.
The first is representation of the LOS observation geometry in a 2° by 2° matrix form,
involving averaging of the spacecraft altitudes, LOS azimuths and declinations, and the
LOS gravity measurements. The solution of this problem is nearly complete. The other
part involves the quantification of the a priori assumptions on the form of a covariance
matrix. The assumption being made is that the topography and compensation are linearly
related, but we don not know the amplitude (i.e. the depth of compensation) of this
relationship. It now appears the best way to deal with this problem is to use a range of
103
depths of compensation for our covariatice matrices, and find the one which minimizes the
chi-squared residuals. The programming for this method is also nearly completed.
Choosing regions of study based on evidence of consistent tectonics has also been a
problem. Division of the surface into radar units based on reflectivity and rms slope (Davis
et al. , 1986, Head et al. , 1985) shows very complex relationships. Since the LOS gravity
and topography show good correlation on a long-wavelength, regional scale, we have
decided to delineate regions based on topography, and will use the residuals of the
admittances to help decide if this delineation is justified. The programs to invert the data
and solve for the equivalent surface mass distribution, and hence the vertical gravity field,
have been completed, as have the programs to determine the admittance functions for the
regions. Theoretical models have been derived for various compensation mechanisms for
comparison with these results. All these programs have been tested with synthetic data and
actual data is now being compiled. Preliminary results are expected shortly, to be followed
by appraisal and interpretation.
Task #4: Contrasts in the Evolutions of Venus and Earth
The analyses related to the Earth's archean are being pursued.
REFERENCES
Aarseth, S.J. , 1972. "Direct Integration of the N-Body Problem," in M. Lecar (ed.).
Gravitational N- Body Problem (Dordrecht, Holland: D. Reidel), 373.
Ahmad, A., and Cohen, L. 1973. "A Numerical Integration Scheme for the N-Body
Gravitational Problem," J. Comp. Phys. , 12,389.
Boss, A. P. 1979. Theoretical Models of Stellar Formation, Ph.D. Thesis, University of
California, Santa Barbara.
Davis, P.A., Kozak,, R.C., and Schaber, G.C. 1986. "Global radar units on Venus
derived from statistical analysis of Pioneer Venus orbiter radar data," J. Geophys.
Res., 91,4979.
Dorman, L.M., and Lewis, B.T.R. 1970. "Experimental isostasy - 1. Theory of the
determination of the Earth's isostatic response to a concentrated load," J. Geophys.
Res. . 75, 3357.
Glatzmaier, G. A. 1984. "Numerical Simulations of Stellar Convective Dynamos. I. The
Model and Method," 3. Comp. Phys. , 5S, 461.
Head, J.W., Peterfreund, A.R., Garvin, J.B., and Zisk, S.H. 1985. "Surface
characteristics of Venus derived from Pioneer Venus altimetry, roughness, and
reflectivity measurements," J. Geophys. Res., 90, 6873.
Jackson, D.D. 1979. "The use of a priori data to resolve non-uniqueness in linear
inversion," Geophys. J. Roy. Astr. Soc. , 57, 137.
Jackson, D.D. , and Matsu'ura, M. 1985. "A Bayesian approach to nonlinear inversion,"
J. Geophys. Res., 90, 581.
Kaula, W.M. 1966. Theory of Satellite Geodesy (Waltham,U&ss.: Blaisdell).
McNutt, M. 1979. "Compensation of oceanic topography: An application of the response
function technique to the Surveyor area," /. Geophys. Res., 84, 7589.
Nacozy, P.E. and Diehl, R.E. 1978. "A Discussion of the Solution for the Motion of
Pluto," Celestial Mechanics, 17, 405.
Wetherill, G. W. 1985. "Occurrence of Giant Impacts During the Growth of the Terrestrial
Planets," Science, 228, 877.
Williams, J.G. and Benson, G.S. 1971. "Resonances in the Neptune-Pluto System, "
Astron. J., 76, 167.
104
ETotatioa of Plasetesmal Velocitiwi^
Glen R. Stewart
University of Virginia
George W. Wetherill
Department of Terrestrial Magnetism
Carnegie Institution of Washington
Abstract
A seif-conaistent set of equations for the velocity evoltitlon of & general plaii»
eteslmal population Is presented. The equations are given m a form convenient for
calculations of the early stages of planetary accumulation, when it is necessary to
model the planetesimad swarm by the methods of gas dynamics, rather than follow
the orbital evolution of individual bodies. To illustrate the relative importamce of
the various terms of these equations, steady state velocities of a simple planetesimal
population, consisting of two diferent sizes of bodies, are calculated. Dynamical
friction is found to be an Important mechanism for transferring kinetic energy from
the larger planetesimals to the smaller ones, providing an ener^ source for the small
planetesimals that Is comparable t© that provided by the vbcous stirring process.
When small planetesimals are relatively abundant, gas drag and inelastic coillslons
among the smaller bodies are of comparable Importance for dissipating energy from
the population.
This research was sponsored by the National Science Foundation under grants AST-82-04256 and
MCS-83-04459 and by NASA under grants NSG-7437, NAGW-398, and NAGW-929.
105
ACCUMULATION OF SOLID BODIES IN THE SOLAR NEBULA
S.J. Weidenschilling and D.R. Davis
Planetary Science Institute, Tucson, AZ 85719
The earliest stage of accumulation of solid bodies in the solar system,
i.e., the formation of planetesimals from dust, occurred in the presence of
the gaseous component of the solar nebula. The effects of the gas on these
particles via aerodynamic drag must be understood in order to produce physi-
cally plausible models of this process. There are major uncertainties in
nebular properties, e.g., its mass (or the gas density), lifetime, and
degree of turbulence. The enigmatic textural and mineralogical properties
of primitive meteorites suggest that we still do not understand the forma-
tion of their parent bodies. Weidenschilling (1980) modeled the settling of
dust to the central plane of the nebula. He argued that coagulation of
grains during settling could be expected due to van der Waals surface
forces, and was probably necessary for the formation of planetesimals.
Coagulation causes settling to be non-homologous; i.e., a small fraction of
the total dust content arrives at the central plane as large aggregates,
while most of the mass remains suspended. Recently Nakagawa et al . ( Icarus
67, 375, 1986) criticized this work because it assumed the dust motions were
always controlled by the gas, whereas the reverse would be true (dust con-
trolling gas) at sufficiently high dust density. However, their conclusion
was based on an analytic homologous settling model. Weidenschilling has
re-examined this problem numerically, using a higher spatial resolution near
the central pleine than previously employed. He found that the first-formed
aggregate layer is thinner than found earlier (~10 km thick at 1 AU for a
-3
low-mass nebula), and containing <10 of the total surface density solids.
The ~meter-sized aggregates are so widely separated that they behave inde-
pendently; i.e., there is no stage at which the dust dominates the motion of
the gas. Earlier, Weidenschilling (1980) had suggested that planetesimal
formation would involve gravitational instability in a layer of >m-sized
aggregate bodies. These new results imply that gravitational instability
might be avoided entirely, due to the extremely low surface density of the
initial layer, with planetesimals forming entirely by collisional coagula-
tion.
An important property of the solar nebula is the non-Keplerian rotation
of the gas due to a radial pressure gradient (Weidenschilling, MNRAS 180,
57, 1977). The resulting headwind" on the planetesimals causes their orbits
to decay. Weidenschilling and Davis ( Icarus 62, 16, 1985) examined the com-
bined effects of gas drag and gravitational perturbations of a planetary
embryo on the orbital evolution of planetesimals. They showed that
planetesimals could be trapped in stable orbits at commensurability reso-
nances (ratio of periods (j+l)/J, where j is an integer). The resonant per-
turbations pump up eccentricities to an equilibrium value e(eq) that is
independent of planetesimal size. The eccentric orbits cause collisional
comminution of the planetesimals. Weidenschilling and Davis suggested a
scenario in which this phenomenon allows an early-formed embryo to prevent
the growth of rivals, while it grows rapidly by capture of small fragments
that are brought through the resonances by drag.
106
Weidenschilling and Ifevis have continued their investigation of reso-
nant planetesimal orbits, using analytic and numerical methods. Three
dimensional orbit integrations have shown that there is no significant pump-
ing of inclinations; the planetesimal swarm remains highly flattened, with a
high collision rate. The rate of eccentricity pumping at a resonance can
be expressed as de/dt 2^ (e{eq)-e)/T, where the time constant t is inde-
pendent of the embryo mass or the resonance order , j . t does depend on the
damping coefficient, i.e., t is proportional to the planetesimal size and
inversely proportional to the gas density. A typical value for a low-mass
3
nebula is t~10 yr for a km-sized body at 1 AU. In the outer nebula, with
7
much lower gas density, t~10 yr for a similar body. Thus, large planetesi-
mals at Neptune's distance may not reach e(eq) before the nebular gas is
dissipated. The presence of fine dust that is coupled to the gas increases
the effective density, and can shorten t. Collisions with other large
planetesimals on Keplerian orbits have no effect on t, but can decrease the
value of e(eq) .
Another important consideration for evaluating the effects of resonant
trapping is the width of a resonance, i.e., the range of semimajor axis (or
mean motion) over which a planetesimal responds to the resonant perturba-
tions. As with any simple damped oscillator, lower damping (a larger
planetesimal or lower gas density) produces a narrower resonance peak near
exact commensurability; higher damping yields a broader range of response
with a lower peak. Using analytic resonance theory (cf . Greenberg, Icarus
33, 62, 1978), it can be shown that the width of a resonance varies slowly
with order j, and is proportional to the square root of the product of the
gas density, the perturbing embryo's mass, and the planetesimal 's size. In
a lower-mass nebula containing an Earth-mass embryo, the effective resonance
-4 -3
width for a km-sized planetesimal is typically ~10 -10 AU. If a planete-
simal has a close encounter or collision with another body, so that its
semimajor axis is changed by more than the resonance width, it may pass
through the resonance without being trapped. Larger bodies have narrower
resonances (less damping), and so require smaller perturbations; also, their
larger t values make it more likely that they will experience a collision or
encounter before their eccentricities can be pumped up. Thus, resonances
are "transparent" to large bodies. This effect alters the accretion
scenario that was proposed by Weidenschilling and Davis. There is a minimum
planetesimal size that allows trapping; below that size drag dominates and
forces bodies through any resonances. The matter accreted by a planetary
embryo is probably bimodal. Much of the mass, especially in the early stage
of accretion, will consist of small (<<1 km) fragments. The late-accreted
mass will include bodies that have grown large enough (>>1 km) to be unaf-
fected by resonances. This scenario can yield rapid growth of the cores of
the giant planets while providing a population of large impactors that pro-
duce their obliquities.
A more detailed description of this work was presented at the sjrmposium
"Origin and Evolution of Planetary and Satellite Systems," Potsdajn, GDR,
October 1986, and has been submitted to Gerlands Beitrage zur Geophysik .
This work was supported by NASA Planetary Geophysics/Geochemistry Program,
Contract NASW-3214.
107
A Scaling Law for Accretion Zone Skes
Yuval Greenzweig and Jack J. Lissauer (University of California at
Santa Barbara)
Current theories of runaway planetary accretion require small ran-
dom velocities of the accreted particles. Two body gravitational ac-
cretion cross sections, which ignore the tidal perturbations of the Sun,
are not valid for the slow encounters which occur at low relative veloci-
ties. Wetherill and Cox {Icarus 63, 290, 1985 ) have studied accretion
cross sections for rocky protoplanets orbiting at 1 AU. Using ana-
lytic methods based on Hill's lunar theory (Petit and Henon, Icarus
66, 536, 1986) one can scale these results for protoplanets that oc-
cupy the same fraction of their Hill sphere as does a rocky body at
lAU. Generalization to bodies of different sizes Is achieved here by
numerical integrations of the three body problem. Starting from Ini-
tial positions far from the accreting body, test particles are allowed to
encounter the body once and the cross section is computed. A power
law is found relating the cross section to the radius of the accreting
body (of fixed mass). The value of the exponent varies with the initial
distribution of inclinations of the test particles. It is found that for an
Initial distribution of planar circular orbits, with uniform semi-major
axes spacing, the cross section obeys an rz power law. For nonpla-
nar circular-orbit distributions, and fixed small Inclination, the cross
sections behave approximately as r. Variable non-zero inclination dis-
tributions result in intermediate power law dependences. These power
laws are valid for the range of parameters of interest to runaway ac-
cretion theories of the planets, but different results apply in satellite
systems near Roche's limit.
108
URANUS AND NEPTUNE: QUESTIONS AND POSSIBLE ANSWERES
Reynolds, R. T. and M. Podolak*, NASA Ames Research Center,
Moffett Field, CA 9^035
Uranus and Neptune are of special interest for theories of the
origin of the solar system, because they represent a special class of
objects intermediate in composition between the giant hydrogen-rich
planets Jupiter and Saturn, and the small, rocky, terrestrial planets.
Their structure and composition should provide important clues to the
origin of the solar system (1,2). In order to compute models of the
internal structure, using high pressure equations of state for the
materials believed to constitute those planets, it is necessary to
measure several parameters characterizing these palnets. Their masses
and radii have been known to sufficient accuracy for many years, and
recently J2 and J^ (respectively the quadrupole and 15-pole moments of
the gravitational field) have been determined for Uranus (3) and moments
Jg for Neptune (4), Until the Voyager flyby, however, the rotation
period for Uranus was not well constrained and various observations
placed it in the neighborhood of 15 hrs. (5). It is interesting that^
although this is a shorter period than the 18 hrs determined
photometrically for Neptune (5), Neptune has the larger J2.
We have constructed a detailed set of theoretical models which
consist of a core of "rock" (MgO, SiO, Fe, and Ni in solar proportions),
surrounded by an envelope of "ices" (H2), CHu and NHo in solar
proportions), surrounded, in turn, by an envelope of these same "ices"
mixed with H2 and He (6). We found although the ratio of "ices" to
"rock" was nearly the same for the two planets, the internal arrangement
was very different in the two cases, a situation very difficult to
explain in terms of current pictures of planet formation. We suggested
(7) that the period measured photometrically for Neptune is influenced
by surface features caused by the motion of Rossby waves in the upper
atmosphere, and , in fact, the body of the planet rotates more
quickly. Such a shorter period is, indeed, indicated by new
measurements of the oblateness (8), The recent Voyager flyby of Uranus
has fixed its rotation period at just over 17 hours, narrowing the
difference between the two models, but by no means eliminating it.
Hopefully, the Voyager encounter with Neptune will help resolve the
issue.
The new improved rotation period, combined with the latest values
for J2 and J^j have raised problems in the understanding of Uranus'
structure independently of a comparison with Neptune. We have found
that none of the models computed previously (fig.1) could be made to fit
with this new set of observational data. Instead it was necessary to
assume that Uranus has an anomalously high ratio of "ices" to "rock",
109
some five times the solar value (9)! Such a high value (fig. 2) may be
a natural outcome of an accretion mechanism we have begun to
investigate. As a planetesimal approaches a protoplanetary atmosphere
with some impact parameter, it may (depending on its composition) lose a
significant amount of mass and be captured by the protoplanet, or lose
very little, and escape to space. Since ice is more volatile, an icy
planetesimal will lose mass at larger impact parameters, and hence the
cross section for capture will be larger for icy planetesimals than for
rocky ones. This provides a possible mechanism for greatly enhancing
the ice to rock ratio over the solar value. This mechanism is currently
being explored.
REFERENCES
1. Podolak, M. and R. T. Reynolds (1984). Consistency tests of theories
from models of Uranus and Neptune. Icarus 5^, 102-111.
2. Podolak, M. and R. T. Reynolds (1985). What have we learned from
modeling giant planet interiors? In Protostars and Planets II
(D. Black and M. S. Matthews, eds.) Univ. Arizona Press pp. 847-
872.
3. French, R. G,, J. L. Elliot, and S. E. Levine (1985). Structure of
the Uranian rings II. Perturbations of the ring orbits and
widths. Icarus 75, 134-163.
4. Harris, A. W. (1984). Physical properties of Neptune and Triton
inferred from properties of Triton. In Uranus and Neptune (J. T.
Bergstralh, ed.) pp. 357*373 NASA CP-2330.
5. Helton, M. J. S. and R. Terrile (1984). Rotational properties of
Uranus and Neptune. In Uranus and Neptune (J. T. Bergstralh,
ed.) pp. 327-347 NASA CP-2330.
6. Podolak, M. and R. T. Reynolds (1981), On the structure and
composition of Uranus and Neptune. Icarus ^6 40-50.
7. Podolak, M., R. Young, R. T, Reynolds (1985). The internal
structures and relative rotation rates of Uranus and Neptune
Icarus 63, 266-271 .
8. Lellouch, E., W, B. Hubbard, S. Sicardy, F. Vilas, and P. Bouchet
(1986). The 1985 August 20 occulation by Neptune: The central
flash determination of Neptune's oblateness and methane
atmospheric mixing ratio. Preprint
9. Podolak, M. and R, T. Reynolds (1986). The rotation rate of Uranus,
its internal structure, and the process of planetary accretion.
Submitted Icarus .
^Permanent address: Dept, of Geophysicaand Planetary Sciences, Tel Aviv
University, Israel.
110
Fig. 1 J4 as a function of the ice/rock ratio -for former Uranus
models. Parameter on right gives enhancement of volatiles
in envelope over solar composition. Tick marks along curve
give rotation period required for fit to J2. Horizontal
lines shoM former limits on observed value of a4.
6r
o
X
2-
_
A =
•
= 75
-
-^
:^
30
^^
.20.
- 1
— ...
-
-
1 1
_ 1
1
1
1
1 1
10
15
20 25
l/R
30 35 40 45
Fig. 2 J4 as a function of the ice/rock ratio for current Uranus
models with 17.24 hr period. Tick marks show enhancement of
volatiles in envelope. Horizontal dashed lines show current
limits on J4.
111
COLLISIONAL AND DYNAMICAL PROCESSES IN M(X)N AND PLANET FORMATION
C.R. Qiapman, D.R. Davis, S.J. Weidenschilling, W.K. Hartmann and D. Spaute
Planetary Science Institute
We report research on a variety of dynamical processes relevant to the
formation of planets, satellites and ring systems. Our main focus has been
on studies of accretionary formation of early protoplanets using a numerical
model, structures and evolution of ring systems and individual bodies within
planetary rings (Ite.vis et al.. Science 224, 1984; Weidenschilling et al . . in
Planetary Rings, p. 67, 1984) and theories of lunar origin (Weidenschilling
et al., in Origin of the Moon, p. 731, 1986; Hartmann, in Origin of the
Moon, p. 579, 1986).
Our earlier work on planetary accretion has been in the context of gas
free accretion. However, a significant area of recent Planetary Science
Institute research concerns the effects of gas drag and resonances on the
accretion of planets. While at present there does not appear to be a strong
requirement for the presence of gas during formation of the terrestrial
planets, there is increasing support for the idea that the giant planets
formed by accumulation of massive solid cores that eventually captured gas
from the solar nebula (Bodenheimer and Pollack, Icarus 67, 391, 1986).
Models for accretion that do not consider affects of gas drag on
planetesimal dynamics are unable to produce such cores on reasonable
timescales.
Weidenschilling and Davis ( Icarus 62, 16,1985) under a separately
funded program showed that orbital decay due to gas drag and resonant
gravitational perturbations lead to stable trapping of planetesimals at
commensurability resonances with a planetary embryo. The synergistic effect
of drag and resonances allows a planetary embryo to pump up eccentricities
of smaller planetesimals to values much larger tham those due to random
_2
encounters. Eccentricities of a few x 10 are induced for plausible
nebular models, independent of the sizes of planetesimals. The close
spacing of resonances ensures that different resonant orbits overlap, and
cross the non-resonant orbits between them. The high relative velocities,
~1 km/sec, cause comminution of the planetesimals. This process could allow
an early-formed planetary embryo to inhibit the growth of potential rivals,
and to dominate the zone covered by overlapping resonances (up to the 2/3
exterior resonances). That scenario alleviates the tendency of numerical
accretion models to produce too many, closely-spaced small planets.
Weidenschilling and Davis noted that if resonances were completely effective
for inhibiting accretion, the spacing of planetary orbits would approximate
their actual values. However, whether this process would yield a single
embryo or several in a given zone depends on details of the size
distribution and its rate of change, impact strengths of planetesimals,
nebular structure, etc. Detailed numerical simulations are needed to
determine the likely range of outcomes.
We are examining in this program the range of outcomes by numerical
simulations of accretion with resonance effects due to an early-formed
planetary embryo. Our approach uses a varient of the method of Spaute et
al., ( Icarus 64, 139. 1985). We have been aided in this effort by
collaboration with Dr. Spaute, who has been in residence at PSI during
1985-86. The simulation computes the outcome of collisions among
112
planetesimals in as many as 50 narrow radial zones spanning a range of
semimajor axis; bodies in different zones interact when their eccentricities
allow their orbits to cross. This spatial resolution allows explicit
inclusion of e-pumping at discrete resonances in specific zones, and the
transfer of mass between zones by collisions and drag. Resonance effects —
de/dt, size limits for trapping, resonance widths, etc. — are parameterized
from results of Weidenschilling and Davis (1985) and Weidenschilling (1986,
submitted to Gerland-Beitrage zur Geophysik).
These simulations involve a determination of the probability density
for collisions between bodies in overlapping orbits, with a Monte Ckrlo
determination of the specific interactions. Collisional outcomes, ranging
from accretion to complete disruption, depend on impact speeds and the
assumed mechanical properties of the planetesimals. In addition, the radial
movement of matter between zones due to secular decay of plane tesimal orbits
is accounted for. A major limitation at present is the lack of detailed
information on the distributions of sizes and orbital elements of the
planetesimals. Due to the complexity of the program, only meein values for
the size, eccentricity, and inclination are computed in each radial zone.
However, these quantities can be weighted (e.g., acording to mass or
cross-sectional area, as appropriate) for assumed power-law size
distributions.
Fig 1 shows the results for one such simulation. An embryo of mass
_7
3x10 Mp. (one Mars mass) is assumed present at 1 AU, and the effects of its
perturbations on a swarm of planetesinra,ls with initial mean radius 1 km is
computed. After a model time of 550 yr, ecentricities have increased
signficantly at resonances. The mean size has decreased sharply at those
locations due to collisional comminution, while accretion occurs in the
relatively quiescent zones between the non-overlapping low-order resonances.
The evolution of eccentricities is dominated by resongmt perturbations, so
any uncertainties in the effects of mutual gravitational stirring of
planetesimals are unimportant to the outcome. Some low-order information on
the size distribution can be obtained from the number of shatttering events
that yield power-law fragment distribution; several tens of percent of the
total swarm has been processed through such events and can be brought to the
embryo for accretion. Future efforts will be directed toward computing more
detailed size distributions.
Hartmann, also working with our visiting scientist, D. Spaute, has
undertaken an investigation of the accretional evolution in a
circumterrestrial (or generally, circumplanetary) cloud of debris. Hartmann
and Spaute visualize the creation of a chaotic, spheroidal or flattened
cloud by an unspecified process (such as giant impact?) in a "stage 1."
They then investigate "stage 2," the collisional evolution that follows from
specified parameters of mean e, i, material properties, etc. in the cloud.
This study used a variant of the accretion numerical model by Spiaute et al.
(1985) which models growth at different distances from a primary. For this
work, Hartmann has updated and modified the treatment of some of the
material parameters based on our experimental results. Preliminary results
113
suggest the very rapid collapse of a chaotic cloud into a flattened ring,
faster growth on the Inner edges of the cloud, and a transition from
fragmentation to accretion as energy dissipation reduces collision
velocities. Thus, if a giant impact threw out a cloud of debris, accretion
might occur first in the inner part of the cloud. A sufficiently large
satellite accreted inside the synchronous point would spiral inward, but
such a satellite accreted just beyond the synchronous point would spiral
outward, sweeping up the rest of the debris. This could account for lunar
formation.
We are currently reviewing our numerical models for calculating
collisional outcomes and orbital stirring due to gravitational encounters.
These algorithms were developed about ten years ago and while they have been
updated occasionally, there has never been a comprehensive review of them in
light of new experimental and theoretical results. We have undertaJcen a
complete review and updating of these models. In particular we will use the
much more extensive collisional database that now exists (e.g. review by
Fujiwara, Mem. Soc. Astron. Ital. 57, No. 1, 47, 1986) and new scaling
relationships (Holsapple and Housen Mem. Soc. Astron. Ital. 57, No. 1, 65,
1986; Housen et al., JGR 88, 2485, 1983) to revise as needed the predictions
of our collisional outcome model. Recent work on gravitational stirring by
Stewart and Wetherill (LPSC XVII Abstracts. 827) 1986. Carusi et al. ( BAAS
18. No. 3, 776. 1986) and Wetherill and Ctox ( Icarus 63. 290. 1985) will be
used to test and improve as needed our present algorithm for modeling this
very important effect for planetary accretion.
4 -
£> 3 -
+J -
1 -
g 0.5
Q_.4 h^„-1\AA4..4/.A AA A ^ A A,
1.05 1.10 1.15
semi-major axis (AU)
1.20
1.25
114
Dynamical Constraints on the Orgin of the Moon
A, P. Boss (DTM Carnegie Institution of Washington) and S. J. Peale
(University of California at Santa Barbara)
Dynamical studies dealing with the orgin of the Moon are de-
scribed and are used to try to eliminate dynamically impossible or
implausible theories of lunar orgin. The orgin of the Moon Is dis-
cussed within the context of the general theory of terrestrial planet
formation by accumulation of planestesimals. The past evolution of
the lunar orbit Is of little use in dlfFerentlatlng between the theories,
primarily because of the inherent uncertainty In a number of model
parameters and assumptions. The various theories that have been
proposed are divided Into six catagorles. Rotational fission and disin-
tegrative capture appear to be dynamically impossible for viscous pro-
toplanets, while precipitation fission (precipitation of Moon-forming
material from a hot, extended primordial atmosphere of volatilized sil-
icates), intact capture, and binary accretion appear to be dynamically
Implausible. Precipitation fission and binary accretion suffer chiefly
from having Insufficient angular momentum to form the Moon, while
Intact capture requires forming the Moon very close to the Earth
without encountering any perturbations prior to capture. The only
mechanism proposed so far that Is apparently not ruled out by dy-
namical constraints and that also seems the most plausible Involves
formation of the Moon following a giant Impact that ejects portions of
the differentiated Earth's mantle and parts of the Impacting body Into
circumterrestrial orbit. The Moon must have accreted subsequently
from this circumterrestrial disk. The giant Impact model contains el-
ements of several of the other models and appears to be dynamically
consistent with the absence of major satellites for the other terrestrial
planets. While the giant Impact mechanism for forming the Moon
thus emerges as the theory with the least number of obvious flaws, it
should be emphasized that the model Is relatively new and has not
been extensively developed nor thoroughly criticized. Much further
work must be done to learn whether the giant Impact mechanism for
lunar formation can be made Into a rigorous theory.
115
LUNAR SCIENCE FROM LUNAR LASER RANGING
J. G. Williams, X X Nawhall , and J. 0. Dickey
Jet Propulsion Laboratory, Pasadena, CA 91109
Seventeen years of lunar laser ranging data have been
analyzed to determine lunar second-degree moment
differences, third-degree gravitational harmonics, Love
number, rotational dissipation, and retroref lector
coordinates. The range accuracy improves through the
time span until 3-5 cm accuracies have been achieved
since 1985. The retroref lectors were located at the
Apollo 11, 14, and 15 sites and the Lunakhod 2 site.
The results from the solution arei
Value Error
Units 10-S
(C-A)/B
531.93
0.76
<B-A)/C
227.95
O.OS
C30
-8.3
0.7
C31
35.2
12.
S31
2,9
3.4
C32
4.812
0.035
S32
1.S82
0,019
C33
l.SS
0.22
S33
-0.30
0.07
A more accurate value of C31 is available from previous
analyses of lunar orbiting satellite Doppler tracking data.
Two additional lunar solution parameters are the potential
Love number, k = 0.027 + O.OOS, and the rotational
dissipation, kT = 0.0048 ± 0.0002 days. The coordinates
of the retroref lectors were determined with accuracies of
a few meters. These results are to be published in the
proceedings of the international symposium titled Figure
and Dynamics of the Earth, Moon, and Planets.
116
SOLAR SYSTEM DYNAMICS
Ja4:k Wkdom, Department of Earth, Atmc^pheric and Planetary Sciences, Massachusetts
Iiistitut® of Technology, Cambridge, Massachusetts 02139
L Rotational Dynamics of Irregularly Shaped Satellites
All irregularly shaped natural satellites tumbled chaotically at the point of capture
into synchronous rotation.
The basic mechanism governing the tidal evolution of the spins and obliquities has
been well understood since the work of George Darwin (1879). If the spin angular velocity
is large the obliquity is driven to a equilibrium value between and ^°. As the spin
is slowed by tidal friction the equilibrium obliquity decreases. If the orbit is fixed, the
equilibrium obliquity goes to zero as the spin angular velocity approaches twice the mean
orbital motion. For smaller angular velocities the obliquity is zero. Spin-orbit resonances
are encountered as the spin rate is reduced and the rotation may or may not be captured
in a spin-orbit resonance with calculable probability (Goldreich and Feale, 1966). ff the
non-synchronous spin-orbit resonances are avoided, the spin continues to decrease until
synchronous rotation is established.
A dramatic exception to the regular tidal evolution envisioned in the standard scenario
is provided by the chaotic tumbling of Saturn's satellite Hyperion (Wisdom, Peale, and
Mignard, 1984). The chaotic tumbling of Hyperion is a consequence or its highly aspherical
shape, to some extent the large orbital eccentricity (e w 0.1), and the fact that the spin
rate of Hyperion has been brought by tidal friction to be nearly synchronous with the
orbital mean miotion.
For other irregulaxly shaped satellites this picture of steady evolution to synchronous
rotation must also be revised. All resonances axe surrounded by a chaotic zone (see
Chirikov, 1979). The width of the chaotic zone surrounding the synchronous spin-orbit
state may be estimated. It turns out to be linearly proportional to the orbital eccentricity,
but exponentially dependent on the parajneter which measures the out of roundness of the
satellite. Thus the chaotic zone surrounding the synchronous state may be significant for
an irregularly shaped satellite even if the eccentricity is not as large as that of Hyperion.
This prediction has been verified numerically. A remarkable fact is that this chaotic zone
seems to always be attitude unstable just as it is for Hyperion. Even Deimos which has an
anomalously low eccentricity of 0.0005 has a non-negligible chaotic zone which is attitude
unstable. In all cases the timescale for the spin «ixis to fall away from the orbit normal
is only a few orbit periods; the attitude instability is very strong. Thus all the irregu-
larly shaped satellites in the solar system, regardless of their orbital eccentricity, tumbled
chaotically at the point of capture into synchronous rotation.
Dissipation within a chaotically tumbling satellite is significantly greater than in a
synchronously rotating satellite. 1£ a secular change in the angular momentum of the
satellite can be ruled out, then the dissipation within the satellite leads to a decrease in
the orbital eccentricity. It is plausible, but not yet proven, that tidal friction will not give
a secular chajige in the angular momentum of a chaotically tumbling satellite because,
speaking qualitatively, the motion is so irregular the tides do not know which way to push.
Once this is rigorously justified then the anomalously low eccentricity of Deimos caji be
explained as resulting from the dissipation in Deimos during the chaotic tumbling phase.
In any case, the chaotic tumbling phase must be taken into account in the orbital histories
of the natural satellites.
This work has been submitted to Icarus.
117
II. Origin of the Kirkwood Gaps
We have used the Digital Orrery (Applegate, et al. 1985) to systematically explore
the phase spa^e of the elliptic restricted three body problem near the principal commen-
surabilities (2/1, 5/2, 3/1, and 3/2). The results for the 3/1 commensurability are in close
agreement with those found earlier with the algebraic mapping method (Wisdom 1981,
1983). Large chaotic zones axe associated with the 3/1, 2/1 and 5/2 resonances, where
there are gaps in the distribution of asteroids. The region near the 3/2 resonance, where
the Hilda group of asteroids is located, is largely devoid of chaotic behavior. Thus there
is a qualitative agreement between the character of the motion and the distribution of
asteroids.
The detailed comparisons (using the representative plane method of Wisdom, 1983)
between the distribution of asteroids and the cheiotic zones do not show perfect agreement.
While the boundaries on the large semimajor «ixis side are in good agreement, there is
in each case a small region at fairly low eccentricity on the small sem^imajor axis side of
the gaps where there are no asteroids and also no chaotic behavior. Unfortunately, the
interpretation of this result is not straightforward. There axe three possibilities: (1) The
planar elliptic approxim.ation is not a sufficiently accurate representation of the motion
near the principal commensurabilities. (2) The void on the representative plane represents
a sufficiently small phase space volume that asteroids were simply unlikely to be found
there. (3) Cosmogonic mechanisms such as resonance sweeping (Henrard and Lemaitre,
1983, and Torbett and Smoluchovski, 1980) have been operative. The first possibility has
been examined by integrating several test particles in the field of the major planets with the
Digital Orrery. Five test particles were started in the region in question and integrated for
five million years each. Of the five test particles three turned out to be chaotic. This gives
strong support to the conclusion that the discrepancy is primarily due to the use of the
elliptic restricted approximation in the systematic exploration. Most likely a combination
of the first two possibilities can fully account for the differences between the systematic
exploration and the observed distribution, but resonance sweeping can not yet be ruled
out. We intend to resolve this question soon.
The existence of chaotic zones at the principal commensurabilities does not by itself
explain the existence of the Kirkwood gaps. We must also understand the mechanism
whereby asteroids on chaotic orbits are removed. We have used the Digital Orrery to
approach this problem as well. Again, five test paxticles were integrated in the field of the
major planets for five million years each. Each one was started in the 2/1 chaotic zone. Of
these five, two reached eccentricities above 0.6, which makes them Mars crossing. In both
cases they went to large eccentricity by a path that temporarily took them to high incli-
nation ($' > 20°). (Recall that the integrations by Froeschle and Scholl (1981) of Giffen's
(1973) 2/1 chaotic orbit seemed to show that there was a maximum eccentricity. Their
integrations were performed in the planar approximation.) This is quite an interesting
result. This is the first example in the solar system where a phenomenon seems to depend
on the interconnectedness of the chaotic zone in many dimensions (the so-called "Arnold
web", see Chirikov, 1979). Sweeping by Mars thus appears to be a plausible removal mech-
anism, though it seems likely that longer integrations will reveal Jupiter itself to be the
culprit.
References
Applegate, J.F., Douglas, M.R., Gdrsel, Y., Hunter, P. Seitz, C. and Sussman, G.J. 1985.
IEEE Trans. Comput. C34, 822.
Chirikov, B.V. 1979. Pbys. Rep. 52, 263-379.
Darwin, G. 1879. Phil. Tr&ns., Part I.
118
FroeschM, C, and SchoU, H. 1981. Astron. Astrophys. 93, 62.
Giifen, R. 1973. Astron. Astrophys. 23, 381.
Goldreich, P. and S.J. Peak 1966. Astron. J. 71, 425.
Henrard, J., and Lemaitre A. 1983. Icarus 55, 482.
Torbett, M. and Smoluchowski, R. 1980. Icarus 44, 722.
Wisdom, J. 1983. Icarus 56, 51.
Wisdom, J., S.J. Feale and F. Mignard 1984. Icarus 58, 137-152.
119
Orbital Resonances J Unusual ConfiguratloBS and Exotk Ro-
tation States among Planetary Satellites
S. J. Peale (University of California at Santa Barbara)
Several examples of satellite dynamics are presented where signif-
icant progress has been made in understanding a complex problem,
where a long-standing problem has finally been solved, where newly
discovered configurations have motivated novel descriptions or where
an entirely new phenomenon has been revealed. The origin of or-
bital resonances is shown in the demonstration of the evolutiona of a
pair oa planetary satellites through a commensurability of the mean
motions by a sequence of diagrams of constant energy curves in a two-
dimensional phase space, where the closed curve corresponding to the
motion in each successive diagram is identified by its adiabatically
conserved area. All of the major features of orbital resonance capture
and evolution can be thus understood with a few simple ideas. Qual-
ifications on the application of the theory to real resonances in the
solar system are presented. The two-body resonances form a basis for
the solution of the problem of origin and evolution of the three body
Laplace resonance among the Galilean satellites of Jupiter. Dissipa-
tion in lo is crucial to the damping of the amplitude of the Laplace
libration to its observed small value. The balance of the effects of
tidal dissipation in lo to that in Jupiter leads to rather tight bounds
on the rate of dissipation of tidal energy in Jupiter. Motion in the
relative horseshoe orbits of Saturn's coorbital satellites is described
very well by a simple expansion about circular reference orbits. The
coorbitals are currently very stable, and their relative motions can be
used for the determination of the masses of both satellites. Pluto and
its relatively large satellite Charon form an unusual system where the
relative size and proximity of Charon lead to a most probable state
where both Pluto and Charon are rotating synchronously with their
orbital motion. The normal tidal evolution of a satellite spin toward
shychronous rotation is frustrated in the case of Saturn's satellite Hy-
perion where gravitational torques on the large permanent asymmetry
cause it to tumble chaotically. Observations of Hyperion's lightcurve
are consistent with the chaotic rotation but do not verify it with cer-
tainty.
120
Dynamics of SatelliteSs Asteroids ^ and Rings
Stanley ¥. Qerm^tt^ CRSi^ Cornell Uniwersity
Work is in progress on: (a) determining the shapes and the internal
structures of satellites (with Peter Thomas of Cornell University); (b)
investigating the tidal heating of Miranda (with Renu Malhotra of Cornell
University and Carl Murray of Queen Mary College, London); (c) investigat-
ing the dynamics of arc-like rings (with Carl Murray); and (d) determining
the structure of the zodiacal cloud as revealed by IRAS (with Philip
Nicholson of Cornell University). Significant progress has been made in
(a) the determination of the shape and the internal structure of Mimas and
(b) understanding the dynamical evolution of Miranda's orbit.
Ca) "Pie Shape and Internal Structure of Mimas
Limb profiles from the six best Voyager images have been used to
determine the shape of the satellite. Correction of image distortions
allows coordinates on the limbs to be located with an accuracy of
approximately one-half picture element: about 0.5 km for the two best
images and between 1 and 2 km for the other images. Ellipses fit to the
limbs show that the shape of Mimas is well -represented by a tri axial
ellipsoid: it is the smallest satellite observed for which this is
possible. The ratio of the differences of the axes , (b - c)/(a-c), is
0.27 ± 0.04, indicating that the satellite is close to hydrostatic
equilibrium. This is the first observation of a satellite in the solar
system with a tri axial equilibrium figure. Using the satellite mass
determined by Kozai (1957) from observations of the libration period and
the libration amplitude of the Mimas-Tethys resonance, and a second-order
theory for the ellipsoidal figure of equilibrium, we deduce that the
satellite has a mean radius <R> of 198.9 ± 0,6 km, a mean density of 1.137
± 0.018 g/cm^ and that the difference between the long and short axes, a -
c, is 17.0 ± 0.7 km. The expected value of a - c for a comparable, but
homogeneous satellite in hydrostatic equilibrium is 20.3 ± 0.3 km. We
conclude that Mimas is probably differentiated and may have a rocky core
of radius ( 0,43 ± 0,10) <R>. The material outside the core probably has
a mean density of 0.98 ± 0.08 g/cm^, consistent with that of uncompressed
water-ice. The rock/ice ratio (by weight) of Mimas is probably a factor
of 2 lower than the cosmic ratio: Mimas is markedly deficient in rock.
This work represents the first determination of the internal structure of
a satellite in the solar system, other than the Moon, and is likely to
shed light on the accretion of satellites. Preliminary considerations
favor the idea of heterogeneous accretion (Dermott and Thomas, 1986).
(b) Tidal Heating of Miranda
This work is part of a continuing effort to understand the dynamics
of the Uranian satellite system. We showed that the theory previously
used to find the masses of the Uranian satellites from their orbital
precession rates contained a fundamental error (Dermott and Nicholson,
121
1986), Thus, we were able to predict that the pre-Voyager masses of the
satellites would prove to be incorrect. This prediction has now been
proved to be right.
We have now analyzed the evolution of the Uranian satellite orbits
due to tidal dissipation in the planet and have calculated the change in
the orbital elements due to passage through low-order orbit-orbit
resonances. We have succeeded in showing that the orbital eccentricity
of Miranda would have been dramatically increased by these passages and
that the subsequent tidal damping of this eccentricity would have heated
the satellite. We have also calculated the change in the orbital
inclination of Miranda on passage through the same set of resonances. The
calculated change is large enough to account for the present very high
orbital inclination of Miranda. Thus, we consider that we have found a
dynamical solution for both the bizarre appearance of Miranda observed by
Voyager and the anomalous orbital inclination (Malhotra, Dermott and
Murray, 1986).
(c) Dynamics of Arc -Like Rings
The location and the stability of the Lagrangian equilibrium points
in the restricted circular three-body problem have been examined under a
variety of drag forces. Linear stability analysis and numerical
integration confirm that, contrary to what might be expected from simple
energy arguments, the L^ and L5 points can be asymptotically stable under
the action of certain drag forces, despite being points of potential
maxima. The results have been extended to the horseshoe regime where the
radial oscillations of the particle are small compared with the width of
the horseshoe path in the rotating reference frame. In this case, the
behavior of the Jacobi constant averaged over the horseshoe path
determines the stability and the sense of evolution of the particle. If
the drag force varies as vP, where v is the velocity of the particle in
the inertial frame, then the value p = 2 is critical for both the tadpole
and the horseshoe regimes. Stability is ensured if sign(a) x sign
(2 - p) is negative.
A similar analysis can be applied to any particle in a co-rotating
arc. These results may have important implications for the stability of
arcs of ring material where the dynamical effects of drag can counteract
the spreading due to particle collisions (Murray and Dermott, 1986),
(d) Structure of the Zodiacal Cloud as Revealed by IRAS
The IRAS Zodiacal History File, which contains the all -sky survey
data, is now to hand at Cornell. Software has been written to Fourier
analyse the data and thereby separate the smooth large-scale zodiacal
background from the narrower dust bands. Our preliminary results were
described in a paper read at the Uppsala Asteroid, Comets and Meteors
meeting (Dermott, Nicholson, and Wolven, 1986).
122
We have previously shown that the dust bands may be debris associated
with the Hirayama asteroid families, in particular, the Eos and Themis
families, and we have predicted (a) that the ecliptic latitudes of the dust
bands should vary with ecliptic longitude and (b) that the central dust
band should be split (Dermott et al,, 1984, and Dermott et al., 1985). We
now have evidence supporting both of these predictions. In particular, we
have evidence showing that the central dust band is indeed split and that
the separation in latitude of the two components is consistent with that
expected for debris derived from the Themis family. This has important
implications for the origin of the particles in the zodiacal cloud.
References
1. Dermott, S. F., Nicholson, P. D,, and Wolven, B, (1986). Preliminary
Analysis of the IRAS Dust Data. In " Asteroids, Comets and Meteors
n^" (eds. C-I. Lagerkvist and H. Rickman) Uppsala, 583-594.
2. Dermott, S. F., Nicholson, P. D., Burns, J. A,, and Houck, J. R. (1985).
An Analysis of IRAS' Solar System Dust Bands. lAU Colloquium No,
85, " Properties and Interaction of Interplanetary Dust" (ed.
Gi ese, R.H. and P. Lamy), D. Reidel Pub. Co., 395-409.
3. Dermott, S. F., Nicholson, P. D., Burns, J. A,, and Houck, J. R. (1984).
Origin of the Solar System Dust Bands discovered by IRAS, Nature,
312 , 505-509.
4. Dermott, S. F. and Nicholson, P. D. (1985). Masses of the Satellite
of Uranus. Nature 319 , 115-120.
5. Dermott, S. F, and Thomas, P. C. (1986). The Shape and Internal
Structure of Mimas. B.A.A.S. 18 ., 761.
6. Malhotra, R., Dermott, S. F., and Murray, C, D. (1986). Tidal Heating
of the Uranian Satellites. B.A.A.S. 18 , 785.
7. Murray, C. D,, and Dermott, S. F. (1986). Dynamical Effects of Drag on
Particles in Corotational Resonances. B.A.A.S. 18, 778.
123
Hamiltonian Theory of Nonlinear Waves in Planetary Rin^'
G. R. Stewart (U. Virginia)
Nonlinear spiral density waves have been observed at several lo^-
tions in Saturn's rings. Our theoretical understanding of these
waves was substantially advanced by the recent derivation by Shu
et al. (1985) and Borderies et ai, (1985) of a nonlinear dispersion
relation which generalizes the linear relation found earlier by Lin
and Shu-^"^ The purpose of this abstract is to report the derivation
of a Hamiltonian field theory for nonlinear density waves. Starting
from the Hamiltonian for a discrete system of gravitating stream-
lines, an averaged Hamiltonian is obtained by successive application
of Lie transforms. The transformation may be carried out to any
desired order in q, where q is the nonlinearity parameter defined In
the above references. Subsequent application of the WKB approxi-
mation yields an asymptotic field Hamiltonian. Both the nonlinear
dispersion relation and the wave action transport equation are easily
derived from the corresponding Lagrangian by the standard varia-
tional principle.
'Work supported in part under NSF grant AST-82-004256 and
NASA grant NAGW-929.
^Shu et aL. Ap. J. 221 (1985) 356
^Borderies et aL. Icarus §3 (1985) 406
124
Planetary Ming Bynamlcs and Morpfielo^
Jeffrey U. Ciizzf ^ Ames Resrarcli Center, principal invretiptor
Richsrd H. Durisen, Indiana University, colwesllgetor
Frank H. Shy, II. C. Berkeley, coiwestlpter
F ring momilet belt paper: We are happy to have finished our long-overdue paper on
the moonlet belt for which we find evidence in the region between Saturn's close-in ringmcwns
Pandora and Prometheus, the so-callKlF ring shepherds. In this paper, Cuzzi and J. Burns show
how the little- noticKl ob^rvations of m^netospheric electron density by Pioneer 1 1 imply
substantial , ongoing injections of mass into the 2000 km region notaj above which surrounds
the F ring. We present a hypothesis that the^ events (which r^iuire the appearance of the ma^
equivalent of a 1 00 m r^ius ice particle every few hours) result naturally from interparticle'
collisions between the sinaller members of an optically thin belt of moonlets. The moonlets in
the belt may have radii ranging from about 1 00 meters to somewhat less than 1 km , and
continually sw^p up the ejecta produced by these collisions, (coping the F region relatively,
but not completely, cirar in st^ly slate. We show that the larpr members of this same
population, numbering perhaps less than a dccen, produce much more massive clumps of
particles when they collide, but that their collisions are far less common. In fact, it is possible
to choose a reasonable parent size distribution that produces simultaneously the many small
events responsible for the Pion^r electron (tensity observations and rare large events that
would lead to rings much like the F ring in mass and structure. These large clumps become
strands when differential motion spreaJs the material longitudinally, and may exist for years
before succumbing to the perpetual swiping &Am of surrounding mronlets.
We thus raise the possibility that the F ring is not necessarily a primordial structure
relying on "shepherding" by Panctora and Prometheus to keep it from spreading, but merely one
of a ^ries of transient features which come and go in response to Kcasional creation and
continuing removal. The mysterious kinks in the F ring would be easily explainaj by (xxasional
clo^ pass^es by moinlets with sizes tea small in general to be visible to the Voy^r cameras,
and its unusually small particle sizes by the fact that the F ring particles are really only a
rasntly lost regalith produced initially by meteoroid bombardment. Even the multistrandaj
nature of the F ring may be explained by this hypothesis, if the regoliths of two rolliding bcxlies
retain some memory of their pre-rallision orbits.
We suspa;t that new insight into similar dynamical puzzles, such as incomplete ring arcs
known to exist in Saturn's Encke gap and in the ring systems of Uranus and Neptune, will be
gained from use of this theoretical framework. It may even help us understand the existence of
the eight Inner rings of Uranus, which seem to have no major shepherds to maintain them from
spreading over the age of the solar system. These and the diffuse dust bands discovered by
Voyager in the inter-^Uranian-ring regions may indicate a situation similar to that which we
describe for the F ring.
The concept of an intimate mix of rings and ringmoons in a highly Interactive situation
just outside Saturn's RtKhe limit, and the increasing evidence for moonlets embedded within
Saturn's main rings (Showalter et al. 1 986), provi* us with new constraints on the properties
of planetary ring systems. These new results seem to increasingly support ^me variation of the
concept first presented by Shoemaker that sizeable moons precede or acxompany ring formation,
and that meta)roid bombardment of such moons, which Is incrrasingly focus^ and intent in
r^ions clo^r to a planet, results in smaller and more numerous mtan fragments closer to the
planet. In aWltion, we suspsit that more detallKl future studies of the dynamical behavior of the
F-r^ion mcanlet belt, which seems to lie outste Saturn's Rxhe limit, may have much to tea^h
125
us about planetary accretion.
First erosion manyscript about half done: We are about halfway through with the
first of an anticipataJ series of papers on transport of mass and momentum by metairoid
ejection in planetary rings. The first paper (somewhat to) rough to be distributed as yet)
provides the theoretical framework of the numerical solution and some Illustrative examples for
certain simple ejajta distributions. It will describe how transport of mass and angular
momentum by both ejecta and viscosity has b^n inducted in the formulation, and what typical
results demonstrate. A companion manuscript will describe our development of the ejecta
distribution in a realistic case Including all known cratering ei&AQ distributions and accurate
distributions of incident metKiroid flux which reflect aberrations due to the orbital motion of
both Saturn and the ring particles, as well as the impact probability which (tepends on Incident
dlr«;tlon and velocity as wel 1 as on the optical depth of the local ring material. Extension of '
certain of these dynamical concepts has been made to demonstrate the Inability of "late heavy
bombardment" impacts to desynchronl2e satellite spins (Lissauer 1 985).
Uranus Ring ring structure and photometry: We are conducting photometric
analysis of the Uranlan rings , working with others of the Yoyac^r imping team. The color of
the rings has been carefully constrained using our im^ analysis system to average hundreds of
pixels along the only barely visible rings. The Improvement in SNR allows ring spectral
contrast to be measured to the 1 5Z level or so; the rings ^em to have the same relatively
colorless spectrum between 0.41 and 0.56 microns as the Uranlan satellites and carbonaceous
chondrites. In a complementary effort, we are analysing the Voyager images over the entire
ran^ of illumination and viewing gsomeXry, using rollatlve transfer modeling to ctetermine the
dust content of the rings and the reflectance of the ring material. The dust content Is extremely
low in all nine main rings. The ring material is about as dark as the darkest primitive
carbonaceous chondrites, and perhaps as dark as Halley. These new results are in agreement
with pre- Voyager estimates reportaJ last year(Cuz2i 1 985). A paper on the ring color has been
submitted to Icarus. One will shortly be written on the dust content and reflectance properties.
Jupiter ring paper finished: We have completed a paper describing our image
processing and analysis of the Jovian ring system (enclosai), and the paper has been accepto!
for publication in Icarus. This work was primarily the thesis re^arch of Mark Showalter ,
although general theoretical collaboration, radiative transfer modeling, and about half of the
ima^ processing were supported or materially aided uncfer this RTOP. This task resulted in a
significantly different view of the structure of the Jovian ring and the discovery of a new
exterior "gossamer" ring, as reported last year,
A ring photometry and structure: Progress continues in photometric and structural
studies of the A ring of Saturn. This work , primarily the thesis research of Luke Danes, is
aimed at determining the distribution of "dust" as a d/namics tracer in and out of spiral density
wavetralns In different regions of the A ring. Our ultimate goal here Is to better understand the
complex process of wave ttemping, which Is less effective in higher optical depth r^ions due,
according to a theory we developed untter this grant , to the lower random velocities therein ( Shu
et a!. 1 985). We expect that these differences will manifest them^lves in the fractional dust
abundance. We have obtained improved constraints on the pha^ function and albedo of the
macroscopic particles in the rings, and soon hope to solve for the dust fraction and Its variation
between unperturbed and perturbal r^lons.
Solution of this problem has been complicatKJ by the poorly constrainaj azimuthally
126
variable brightness shown by the A ring; that is, in orcfer to determine a pha^ curve that can
constrain the fractional abundance of dust, we n^ first to ctetermine how certain of the
brightness observations may be corruptaj by the ^imuthal variation, which ^ems to have an
amplitufte of as much as AQ% in certain cases and is pcwrly constrained as to how the phase of
the effect chan^ with viewing gK)metry. We continue to spend a reasonable amount of effort
trying to understand the systematics of this effect, at least to the extent possible to avoid having
it bias our primary results.
Improvements to Image processing system: The powerful and flexible image
prccessing system which we cteveloped, support, and continually upgracte under this RTOP
provides a major fasility for about ten researchers. We hope to maintain the ability to support
both resident and visiting scientists interested in ring studies, and to be able to provide analysis
in support of research tastes initiated by others than imagino team members. For instanre. we '
recently were able to rapidly address questions posed by members of the PPS team on some
apparent anomalies In the Uranian ring structure, and by a member of the Electric Field
Analyser team on the possibility of ctetecting extended prions of dust at the location of Voyager
ring plane crossing. Most recently, we have implemented on- line inspection and display
programs for axBSsing the full stellar and raiio occultation data sets. We actively pursue the
use of computer telecommunications and data transfer between interested ring scientists. The
programmer/analyst time required to maintain this system and support occasional users, as
well as to actually perform much of the basic data analysis, is funded partially under this RTOP
and is our fundamental limitation. For the sake of economy, we will continue to run
interactively on the Ames Space Science Division VAX'es ( 1 1 /785 and 8600) for the near
future.
Bibliography:
Cu2z1, J. N. (1985) Rings of Uranus: not so thick, not so black, Icarus, 63, 312-316
Cuzzi, J. N. and J. A. Burns ( 1 986) Charged Particle depletions surrounding Saturn's F ring -
evidence for a moonlet belt; Icarus, submitted
Lissauer, J. J. (1 985) Can cometary bombardment disrupt synchronous rotation of planetary
^tellites? J. Geophys. Res., 90, 1 1289- 1 1293
Ockert, M. E. , J. N. Cuzzi , and C. C. Porco ( 1 986) Uranian rings: photometry from Voyager;
paper pre^nted at Paris DPS mating, October 1 986
Porco, C. C. , J. N. Cuzzi , M. E. Ockert, and R. J. Terrile ( 1 986) The color of the Uranian rings;
Icarus (submitted)
Showalter , M. , J. N. Cuzzi , E. A. Marouf , and L W. Esposito ( 1 986), Satellite wakes and the
orbit of the Encke gap moonlet; Icarus, 66, 297-323
Showalter. M. , J. A. Burns. J. N. Cuzzi, and J. B. Pollack ( 1 986). Jupiter's ring: new results
m structure and particle properties, Icarus, In press
Shu, F. H., L. Dones, J. J. Lissauer, C. Yuan, and J. N. Cuzzi (1985), Nonlinear spiral density
waves: viscous damping, Astrophys. J., 299, 542-573
127
Dynamical Studies of Satum^s Rings
Philip D. Nicliolsony Astronomy Department, Cornell University,
and Carolyn C. Porco^ Lunar and Planetary Laboratory^
University of Arizona.
We are currently pursuing several investigations of Saturn's
rings, employing data from three Voyager experiments; radio
science (RSS) , stellar occultation (PPS) , and imaging. These
investigations, which are concentrated on the poorly-
understood but regularly-organized C ring, include a search
for eccentric or inclined features? a photometric study of
regions of different optical depth; an analysis of wavelike
structures at three locations; and a study of the size-
distribution of meter-sized particles (with M. Showalter of
NASA-Ames Research Center) . Recent results from these
investigations are summarized here.
a) Eccentric features in the C ring.
Of 35 well-defined features examined in the outer C ring,
only 4 show rms deviations from circularity in excess of 2 .
km. These features form the inner and outer edges of two
relatively opaque ringlets which lie in, or adjacent to, two
narrow gaps at radii of 1.470 R_ and 1.495 R_ (1 R_ = 60,33
km). The 16 km wide 1.470 R ringlet, with a radial amplitude
of ±2.2 km, appears to owe its eccentricity to perturbations
by the 2 t 1 inner Lindblad resonance with the satellite
Prometheus, although it is possible that the ringlet is freely
precessing instead (Porco and Nicholson, 1985) .
The 64 km wide ringlet at 1.495 R , on the other hand, is
almost certainly a freely precessing ellipse, similar to the
previously studied Maxwell ringlet (Porco et al., 1984). Its
radial amplitude is ±3.9 km. These results contradict the
expectation that the edges of these two ringlets would be
perturbed by two nearby Mimas 3:1 resonances, and call into
question the role played by these resonances in forming the
surrounding gaps.
(b) WaveliJce structures in the C ring.
Much of the fine scale structure in the A ring, and a
smaller portion of that in the B ring, has been shown to be
due to density waves and bending waves driven at resonances
with external satellites (e.g., Holberg 1982; Esposito et al.,
1983) . Perhaps because of the weaker resonances in this
region, no such waves have yet been identified in the C ring.
There are, however, three wavelike features associated with
narrow gaps at 1.28 R , 1.470 R , and 1.495 R , We have
128
examined these waves in both the RSS and PPS data, and
attempted to fit both density wave models and models
describing perturbations by moonlets orbiting within the gaps
to the observations.
Rather surprisingly, no model has been found which
satisfactorily accounts for either the 1.470 or 1.495 R^
waves, although 14-20 complete oscillations are observed. The
wave at 1.28 R , with ~ 15 oscillations, has the character-
s'
istics of a density wave, but the associated resonance appears
to be far too weak to account for the wave's observed
amplitude. Furthermore, the gap associated with the wave,
with a width of 15 km, appears only in the RSS occultation
data, suggesting either temporal or longitudinal variability,
or extreme particle size segregation.
(c) Particle size distributions
We have shown (Showalter and Nicholson, 1986) that the
statistical properties of the Voyager PPS stellar occultation
data are incompatible with simple photon (i.e., Poisson)
statistics. An improved statistical model has been developed,
which takes into account the random distribution of finite-
sized particles within the rings. When applied to the PPS
data, this model yields different, but internally consistent,
results for the 4 principal ring regions. For power law size
distributions of plausible slopes (d(log n)/d(log r) > - 5) ,
the maximum particle size is found to be r(max) < 3m in the C
ring, < 5m in the Cassini Division, and ~10m throughout the A
ring. The B ring shows a gradient in r(max) , from ~ 5m in the
inner region to 10-15m in the outermost region. Our results
are consistent with average size distributions obtained for
the C and A rings by Zebker et al. (1985), based on scattering
cross sections at radio wavelengths, but offer the first
detailed insight into variations in the size distribution.
References :
Elliot, J.L. and Nicholson. P.D. (1984) . In Planetary Rings ,
R. Greenberg and A. Brahic, Eds., Univ. of Arizona Press.
Esposito, L.W. , O'Callaghan, M., and West, R.A. (1983).
Icarus 56 , 439.
Holberg, J.B. (1982). Astron. J. 87 . 1416.
Porco, C.C., and Nicholson, P.D. (1985). BAAS 17 , 716.
Porco, C.C., Nicholson, P.D., Borderies, N,, Danielson, G.E.,
Goldreich, P., Holberg, J.B., and Lane, A.L. (1984).
Icarus 60, 1.
129
CHAPTER 5
PLANETARY INTERIORS, PETROLOGY, AND GEOCHEMISTRY
HIGH PRESSURE COS MO CHEMISTRY OF MAJOR PLANETARY INTERIORS: LABORATORY
STUDIES OF THE WATER-RICH REGION OF THE SYSTEM AMMONIA-WATER
Malcolm Nicolj Mary Johnsons Steven Boonej and Hyunchee Cynn
Department of Chemistry and Biochemistry
University of Californiaj Los Angeles, CA 90024
Models of the major planets and their satellites make simple, rather
arbitrary assumptions concerning deep interiors. "Rock" cores or "ice"
(water-ammonia-methane) layers are often invoked without considering
thermodynamic consistency* The behavior of "gas-ice" mixtures at very high
pressures, however, is poorly understood. When this project began, few
measurements existed on binary or multicomponent gas-ice systems at
pressures of the order of 1 GPa. We set out, therefore, to determine some
relevant pressure-temperature-composition (P-T-X) regions of the
hydrogen (H2) - helium (He) - water (H2O) - ammonia (NH3) - methane (CH4)
phase diagram. These experiments, and theoretical modeling of the relevant
phases, are needed to interpret ice-gas systems of planetary interest.
Our first goal was to show that the data needed to characterize
compositions and structures of a multi-phase, multi-component system at
very high pressures can be obtained with reasonable precision. We began
with water-rich solutions of ammonia for several reasons. These mixtures
are relevant to planetary interiors and are relatively easy to prepare. A
few parts of the relevant P-T-X space had been determined. These include:
the P-T surfaces of pure water [see, for example, Mishima and Endo ,1980]
and pure ammonia [Mills et al., 1982] and the T-X surface of
ammonia-water at atmospheric pressure [Rollet and Vuillard, 1956] which
includes two water-rich compounds ammonia dihydrate (NH3*2H20) and ammonia
hydrate (NH3''H20). The dihydrate is the ammonia-bearing phase most likely
to occur near the surfaces of icy planets. Recent models [Lunine and
Stevenson, personal communication] suggest that both dihydrate and hydrate
are important in the evolution of the interior and surfaces of the icy
satellites. The work was begun by Ms. Andrea Koumvakalis and Dr. Mary
Johnson [Johnson et al, , 1985] and is being continued by Mr. Steven Boone
and Mr. Hyunchee Cynn.
Even for this relatively simple system, many experimental problems
had to be overcome. The structures of some of the solids are very
different from the structure of the liquid. These solids are difficult to
grow, and liquid or a glassy solid often persists metastably at conditions
where crystals should form. Ice VI and ammonia dihydrate are very
troublesome. Despite many attempts to overcome this metastability, it is
not clear that equilibrium boundaries have been established among these
phases and the liquid. All of the phases also are colorless; and the
structures and vibrational spectra of many of the phases are not known.
Thus, we learned to identify phases by shapes and birefringence. These
observations do not identify phases unambiguously; however, they help us
us to determine how to grow authentic samples of pure phases.
A manuscript submitted to the Journal of Geophysical lesearcb during
August 1986 describes Dr. Johnson's studies of the phase diagram of
133
(NH3)x(H20)i_x at pressures to 5 GPa, temperatures from 240 to 370 K, and
ammonia compositions to 50%. Particularly careful studies were made near
5j lOj ISj 20s 25j 30j and 34% in an effort to resolve questions about
metastability of the liquid and possible differences between visual work
in the diamond-cell and thermophysical measurements in a large
piston-cylinder apparatus. Dr. Johnson showed that melting at 20% NH3 and
higher compositions is complex; the composition of the high-temperature
solid phase varies with both temperature and time; and the phase diagram
can be reasonably well constrained by these data. Other results obtained
from Dr, Johnson's work are:
1, At 25° C and lower temperatures, five well characterized and
well behaved phases were observed, including: liquid; ice VI; ice
VII; an ammonia hydrate; and an ammonia dihydrate.
2, At 3.45 GPa, dihydrate reconstitutes into myrmekitic intergrowths
of Ice VII with a higher-relief phase which seems to be isotropic.
The high-pressure phase was shown to have a composition near
NH3''H20, The transition is either independent of temperature or
has a negative P-T slope. Hydrate plus Ice VII remains stable at
room temperature to 14.7 GPa. Whether the low-temperature and
high-pressure hydrates are equivalent must still be examined.
3, To the limits of precision of this study, none of the boundaries
between solid phases of water ice are shifted by the presence of
ammonia, although melting is suppressed in the expected manner.
4, At room temperature, the Ice Vl-dihydrate- liquid eutectic is at 17
atom percent ammonia and 1.58 GPa. The eutectic curve as a whole
may be fit by the equations: T = 174 + 1.125 P - 0.0229 P^ (r =
0.83) and X = 13.35 + 0.0252 P (r = 0,96). 5, The dihydrate- liquid
field has an unusually-shaped region of stability. Dihydrate
becomes very prominent in the high pressure regieme. At room
temperature, dihydrate melts at 0.88 GPa.
Mr. Steven Boone is extending the visual studies system in the
diamond-anvil high pressure cells, especially near 33% NH3.
Mr. Hyunchee Cynn used the data obtained by Dr. Johnson to grow single
crystals of ammonia dihydrate at room temperature and has begun to determine
their structures by x-ray diffraction. Preliminary data suggest that
dihydrate has a monoclinic-B structure. The lattice constants were:
a = 709.80 pm, b = 568.86 pm, c = 886.64 pm, a = 90.008°, 6 = 109.545°, and
Y = 90.401°. While the diffraction work continues, Mr. Cynn also has grown
single crystals of dihydrate for Raman and infrared spectroscopy.
Drs. Johnson and Nicol are involved in two other projects of planetary
interest. (1) Chemical reactions during shock compression of simple
molecules are being followed by spectroscopy in order to evaluate how the
reactions affect the interpretation of equation of state data obtained by
shock methods. (2) Temperature and x-ray diffraction measurements are being
134
made on resistively heating wires in diamond-anvil cells in order to obtain
phase and structural data relevant to interiors of terrestrial planets.
Shock compression data are the other major source of equations of state
for H-C-N~0 compounds at pressures and temperatures inside planets. However,
many compounds undergo chemical reactions during shock compression. The
products and kinetics of these reactions are not understood, and the effects
of the reactions have not been fully considered in reducing shock wave data
to equations of state. With benzene as a prototype, we use molecular
emission spectra to detect products of these shock reactions at pressures
between 20 and 65 GPa and temperatures between 2000 and 5000 K. [Johnson et
al., 1986] The spectra show many bands of C2 and other small molecules.
Recent experiments with doubly-shocked material suggest that kinetics of
the post-shock reactions can be followed.
Another shock wave project involves a "synthetic Uranus", a solution
of iso-propanol, ammonia, and water with C:N:0 compositions of cosmic
abundance. In addition to the spectroscopy, shock equations of state and
electrical conductivities of this "planet" are being measured. Other
starting mixtures also will be studied in order to determine whether the
results depend strongly upon the initial chemical species.
Drs, Boehler, Johnson, and Nicol also developed techniques for
measuring the temperatuers of resistively-heated iron and other metals in
gasketed diamond-anvil cells. [Boehler et al. , 1986] Dr. Boehler has applied
these techniques to determination of the phase diagram of iron and, with
collaborators from Paris VI, Paris-Nord, Riso, and HASYLAB, they are
obtaining high-pressure, high-temperature x-ray diffraction data for alpha,
gamma, and epsilon iron that are needed to understand the nature of the
Earth's core.
REFERENCES
Boehler, R. , M. Nicol, C.-S. Zha, and M.L, Johnson, 1986, Physica B 139/140,
916-918.
Johnson, M.L. , M, Nicol, and N.C. Holmes, 1986, in Y.M. Gupta, ed., Shock
Waves in Goadensed Matter (Plenimi, New York) 201-206.
Johnson, M. L. , A. Schwake, and M. Nicol, 1985, in Klinger J. et al,, eds,.
Ices in tlse Solar System (D. Reidel, Doerdrecht) 39-47,
Mills, R.L., D.H. Liebenberg, and Ph. Pruzan, Ph. (1982), J. Phys. Che®, 86,
5219,
Mishima, 0,, and S.J, Endo, 1980, J, Chem. Phys. 73, 2454,
Rollet, A. -P., and G. Vuillard, 1956, Ctmptes. rendns Acad. Sci. Paris 243,
383.
135
iaCPElilMENTS FEOTAMMG TO THE FORMATIOM AND
EQUIUBRATION OF PLANETARY COKES
Principal Investigator: Raymond Jeanloz
Dept. Geology and Geophysics
University of California
Berkeley. CA 94720
I
High- Pressure Metallization of FeO
and Implications for the Earth 's Core
Elise Knittle and Raymond Jeanloz
The phase diagram of FeO has been experimentally determined to pressuiresa)f
155 GPa and tem.peratures of 4000 K using shock-wave and diamond-cell tecii-
niques. We have discovered a metallic phase of FeO at pressures greater than ?0
GPa and temperatures exceeding 1000 K. The metallization of FeO at high pres-
sures implies that oxygen can be present as the light alloying element of the
Earth's outer core, in accord with the geochemical predictions of RingwQod. The
high pressures necessary for this metallization suggest that the core has acguired
its composition well after the initial stages of the Earth's accretion. Direct experi-
mental observations at elevated pressures and temperatures indicate that core-
forming alloy can react chemically with oxides such as those forming the mantle.
The core and mantle may never have, reached complete chemical equilibrium, how-
ever. If this is the case, the core-mantle boundary is likely to be a zone of active
chemical reactions.
coo 2000
TEWtRATURE OO
KD^^
K?-
i Kf-
c-
Dianxxid-C«ll
Dato
^%.,fi
Shock- Wave 5
Data
20 ' 40 ' eD ' 80 ' 100' lio' Mo' 160
Pressure (GPa)
Figure 1. Phase diagram of PeO based on high-
pressoire experiments witli the diamond cell (DAC)
and by shock- wave techniques (Hugoniot).
Diamond-cell experiments have been carried out
both at room temperature (300 K) and at com-
bined high temperatures £ind pressures, as indi-
cated on the figure. The phase boundaries
between antiferromagnetic (nonmetallic),
paramagnetic (nonmetallic), and m.etallic phases
are shown by heavy lines and a shaded region; the
latter reflects the imcertainty in the slope of the
boundary. The soMus is given by a dashed line.
Figure 2. Room-temperature diamond-cell (0 to
83 GPa) and high- temperature shock- wave (72 to
155 GPa) electrical resistivity data for FeO.
Shock-wave resistivity measurements for iron
[Keeler and Mitchell, 1969] are summarized (solid
hne) to illustrate that the resistivity of FeO is me-
tallic at elevated pressures and temperatiires.
The three order of magnitude difference in resis-
tivity between the diamond-cell and shock-wave
measurements at 80 GPa cannot be explained as a
temperature eflect on semiconducting FeO. The
dashed line is the calculated resistivity for sem-
iconducting FeO at high temperature. At 72 GPa
and 1200 K (the calculated Hugoniot temperature)
the predicted resistivity is still two orders of mag-
nitude higher than that observed in the shock-
wave experiments.
136
II
Ihe Melting Curve of Iron to Over 100 GPa
und of Iron Sulfide to Over 60 GPa
Quentin Williams and Raymond Jeanloz
We have measured the melting curve of Fe and FeS to pressures exceeding 100
GPa and 60 GPa, respectively, using a laser-heated diamond cell. Temperatures of
the samples were determined spectroradiometrically, and melting was determined
based upon a combination of visual observation during heating and of textur^l
observations on the quenched samples. To avoid chemical reaction with the dia-
mond anvils, all samples were suspended in an AlgOs (ruby) matrix. Our results for
iron agree with previous measurements below 7 GPa, and indicate a melting tem-
perature of 4100 (±200) K at 100 GPa. By extrapolation, we find that the melting
tera.perature of iron at the pressure of the core-mantle boundary (136 GPa) is 4500
(±300) K. Our data for FeS suggest an initially steeply rising melting curve (tem-
perature of 3000 K at 50 GPa), with greater curvature at high pressure than that
observed for iron. We find no evidence for the hot iron samples reacting with the
ruby when they are solid, but both molten Fe and molten FeS appear to react
significantly with the oxide at elevated pressures.
5CXX)
4000
;3000
■"^K
1000
\m
Jf
Ay^'^^
Y/
,^Stiong
etol.
I I I I I I
Solid
Core-Montle
Boundary -
Iron
I I I
50 100
Pressure (0>d
Pigiire 3. Bounds on the melting curve of iron
from static experiments as a fimction of pressure.
Solid triangles represent the highest temperature
measured on solid iron samples at a given pres-
sure, -while open triangles indicate the lowest tem-
perature of Uguid samples. Lengths of symbols
reflect the statistical uncertainty in temperature
from the relevant spectral fit. The low-pressure
melting curve of Strong et al. (2) is also shown,
along with the phase equilihria of the known iron
crystalline phases: a represents the body-
eentered-coibic structure, e the hexagonal close-
packed structure, and y the face-centered cubic
structure.
137
Melting of Troilite at High Pressure in a Diamond Cell by
Laser Heating
William A, Bassett and Maura S. Weathers
Dept. Geological Sci», Cornell U,, Ithaca, N.Y, 14853
We have assembled a system which allows us to measure
melting temperatures at high pressures. Figure 1 is a schematic
diagram of our apparatus. The sample is heated with radiation
from a YAG laser ( A = 1,06 um) , either continuously or in
a Q-switched mode. The Q-switch can be used to produce
individual pulses of 100-200 nsec duration. The power of
each pulse is of the order of 1,000 watts. The laser can
be triggered by the computer (Fig. 1) at the same time that
the photodiode array starts to collect a series of spectra.
The beam is reflected downward through a microscope objective,
through the upper diamond anvil, and focused onto the sample.
The laser light is thereby strongly converging onto the
sample, producing intense heating only at the sample and
not within the diamond anvils. The heated area is "10 um
in diameter. A vidicon system is used to observe the sample
during heating.
Incandescent light from the heated sample passes back
through the objective lens into a grating spectrometer.
The spectrum of the incandescent light is received by the
photodiode array and stored in the multichannel analyzer.
These data can then be transferred to floppy disk for analysis.
A curve-fitting program is used to compare the spectra with
standard blackbody curves and to determine the temperature.
Pressure is measured by the ruby fluorescence method,
A frequency-doubling crystal is placed in the laser beam
to produce green light having a wavelength of 530 nm. This
light, when focused onto particles of ruby inside the diamond
cell, excite fluorescence. The wavelength of the emitted
red light can then be used to measure the pressure. The
pressure is measured at room temperature by this method.
We have tested our apparatus by melting several different
materials. These materials provide various criteria for
determining whether or not melting has occurred. Our technique
in these experiments is to bracket the melting temperatures
of the materials. During each laser pulse a spectrum of
the emitted incandescent light is collected and stored.
It is necessary to remove the sample and look for evidence
of melting using optical and electron microscopy. We have
used various criteria in different samples to establish
whether or not melting has occurred, including the formation
138
from elongate fibers of platinum, formation
from angular particles of silicon and diamond,
a groove with ridges on a diamond face, encap-
droplets of the pressure medium within diamond
1984), and holes melted through tungsten foil.
We have also inferred melting from electron diffraction
patterns of samples (Weathers and Bassett, 1986),
of droplets
of droplets
formation of
sulation of
(Gold et al,
The samples which are loaded into the diamond anvil
cells consist of a mixture of small grains (<10 microns)
of the actual sample (e.g., graphite, troilite, diamond,
iron, etc.) and a pressure medium (e.g., LiF, NaCl , or KBr).
Usually between 20 and 50 grains in each sample are heated
using the laser radiation. The position of each heated
grain is recorded on a photograph. When the sample is removed
from the diamond cell, the pressure medium is slowly dissolved
in alcohol so that the grains maintain their relative positions
and can be identified from the earlier photographs. The
dissolution of the pressure medium is done so that the sample
is deposited on a grid which can be put directly into the
electron microscope.
We are studying the melting behavior of natural troilite
(FeS). Powdered troilite mixed with NaCl at "200 kbar was
laser-heated. Our initial transmission electron microscopy
study showed that we successfully melted grains of the troilite,
converting angular particles into spherical grains (Fig, 2),
An example of a preliminary temperature determination is
shown in Figure 3. A blackbody curve for 2800 K was fitted
to the processed spectrum which represents a grain of troilite
that melted at 5 kbar. Further analyses of temperatures
obtained at various pressures, combined with determinations
of whether or not the heated grains melted will allow us
to bracket the melt curve for troilite.
Gold, J.S., Bassett, W.A,, Weathers, M.S., and Bird, J,M, (1984)
Melting of diamond, Science 225, 921-922,
Weathers, M.S., and Bassett, W.A., Melting of carbon at
300 kbar, submitted, Phys , Chem. Min.
139
TfcQ=| - f y P
JtrnNIUTOR / O- SWITCH
VAQ LASER
'^DMHOND ANVIL CtLL
Figure 1 . Diagram showing our experimental
apparatus for melting materials at high
pressures and determining temperatures.
The YAG laser can be rapidly pulsed
using the Q-switch. The intensity of
the laser light is controlled by means
of an attenuator. The laser beam is
focused onto the sample in the diamond
cell. A beam-splitter mirror allows
a portion of the light to be used to
monitor the sample during the experiment.
The incandescent light given off by
a heated sample is directed to the grating
spectrometer and photodiode array.
Spectra collected by the photodiode
array are stored in the computer controlled
optical multichannel anlyzer and are
later processed to determine temperature.
Figure 2, Transmission electron
image of spherical grains of
troilite (FeS). These grains,
which were originally angular,
were melted by laser-heating.
2800 K
SOO
-I — 1 — 1 — I — r
550
-I — I — 1 1 r — 1 — I — r
SOO 650
wavelength (ran)
i t I I I
700
750
Figure 3. Processed spectrum
from a heated grain of troilite
fit to a blackbody curve of
2800 K. The pressure at the
time of heating was 5 kbar.
This spectrum is from a grain
that showed evidence of melting.
140
PROPERTIES OF PLANETARY FLUIDS
AT HIGH PRESSURE AND TEMPERATURE
W. J. NelUs, 0. C. Hamilton, N. C. Holmes^ H. B, Radousky,
F. H. Ree, M. Ross, D. A. Young, and M. Nicol*
Lawrence Livermore National Laboratory, University of California
Livermore, CA 94550
*University of California, Los Angeles, CA 90024
The outer planets Uranus and Neptune are thought to consist of the
"ices" H2O, CH4, NH3, and possibly CO, CO2 and N2-' The envelopes of
these planets, as well as those of Jupiter and Saturn, are composed of
H2 and He J In order to derive models of the interiors of these planets
we have been studying the equations of state and electrical
conductivities of these molecules at high dynamic pressures and
temperatures. This study is timely because of the recent Voyager II
flyby of Uranus, which measured the magnetic field distribution and the
rotation rate of the magnetic field, which gives the rate of rotation of
the interior of the planet. 2 The gravitational moments derived from the
rotation rate put constraints on the mass distribution. Thus, equations
of state of representative materials are needed to model the chemical
composition. The magnetic field requires information on the electrical
conductivities of representative materials to develop dynamo models.
The condensed gases in these planets were compressed isentropically
starting from very low density and temperature. By virtue of their
large masses, however, interior temperatures and pressures are quite
large. For Uranus the conditions in the "ice" layer range between
0.2-6 Mbar and 2000-7000 K.3 Shock compression of liquids achieves
virtually the same states in the laboratory.
Interest has been generated in the interior of Uranus because of
its unusual magnetic field with an offset tilted dipole moment of
0.23 gauss Ry-^, whose dipole axis is tilted about 60° from the axis of
rotation and centered at about 0.3 R^, where Ry is the radius of
Uranus. 2 Thus, material at substantially larger radii than 0.3 Ry must
be contributing to the dynamo. This means that data we obtain using our
two-stage light-gas gun at pressures up to ~ 2 Mbar and temperatures up
to 5000 K are especially relevant for constructing models of the
interior.
During the past year we used the fast optical pyrometer developed
the previous year to complete shock temperature measurements for N2 and
CH4. Nitrogen can exist inside Uranus by virtue of the decomposition of
NH3, whose shock temperature we measured previously.^ The nitrogen
temperature data showed that it undergoes a continuous phase transition
to a dense, stiff, monatomic, diamond-like phase above 0.3 Mbar and
6000 K. The temperature data allowed us to demonstrate shock-induced
cooling (a first), that (sT/aP)y < in the transition region (as
predicted), and the existence of crossing isotherms in P-V space. 5 The
shock-induced cooling is caused by absorption of internal energy by
dissociation. Since N2 at room temperature does not dissociate at
static pressures up to 1.3 Mbar,5 the transition in the shock-wave
141
experiments must be temperature-driven. Because of the high
temperatures inside Uranus, the same phenomena would be expected there
also. Thus, it is quite likely that the lower ice region is composed of
stiff, diamond-like H, C, N, phases. In contrast, our two shock
temperature points near 0.4 Mbar and 4000 K for liquid CH4 are in good
agreement with the published prediction, which assumes shocked methane
is in the molecular phase. ^
Electrical conductivities measured for shocked liquid CH4 show that
its conductivity is substantially lower than that of H2O. Our results
for N2 show that our maximum observed value at 0.6 Mbar is comparable to
the maximum conductivity we measured for H2O, but the conductivity of
H2O is much higher than for N2 below 0.6 Mbar.
In order to find materials with higher electrical conductivities
than observed so far for individual fluids and which may be necessary to
derive a Uranian dynamo theory, we have started to investigate fluid
mixtures. We have devised a mixture we call "synthetic Uranus," which
is based on estimates of the composition of Uranus, proposed to be
mostly H2O and H2. We have made a liquid with a 4:1 ratio of H to 0,
corresponding to an equimolar mixture of H2O and H2, and also with
concentrations of C and N such that the ratios of the concentrations of
to C and to N correspond to the ratio of their cosmic abundances
(7:4 for to C and 7:1 for to N).^ Three shock-wave equation-of-
state points were measured between 0.15 and 0.78 Mbar. This mixture has
the same shock P-V curve as H2O when scaled to the appropriate initial
molar volume. This fluid has a relatively high conductivity at
ambient. Our preliminary data does not allow us to say that its
conductivity is larger than that of water at higher pressures and
temperatures. Work is continuing on this point.
Our research plans are to complete our measurements of shock
temperatures, emission spectroscopy, and electrical conductivities of
the "ices," H2, and "synthetic Uranus" and to develop techniques to
measure the same properties in quasi-isentropically compressed fluids.
This technique would cause the specimens to track states closer to the
planetary isentropes; that is, at relatively higher densities and lower
temperatures than the shock data.
Work performed under the auspices of the U.S. Department of Energy
by the Lawrence Livermore National Laboratory under contract number
W-7405-ENG-48.
REFERENCES
1. 0. J. Stevenson, "Interiors of the Giant Planets," Annu. Rev. Earth
Planet. Sci. TO, 257 (1982).
2. N. F. Ness, M. H. Acuna, K. W. Behannon, L. F. Burlaga, 0. E. P.
Connerney, R. P. Lepping, and F. M. Neubauer, "Magnetic Fields at
Uranus," Science 233» 85 (1986).
3. W. B. Hubbard and J. M. MacFarlane, "Structure and Evolution of
Uranus and Neptune," J. Geophys. Res. 85, 225 (1980),
142
4. H. B. Radousky, A. C. Hitchell, W. 3. Nellis, and M. Ross, "Shock
Temperature Measurements in Anroonia/' in Shock Waves in Condensed
Matter, edited by Y. M. Gupta (Plenum, New York, 1986), pp. 467-472,
5. H. B. Radousky, W. 0. Nellis, M. Ross, D. C. Hamilton, and A. C.
Mitchell, "Molecular Dissociation and Shock-Induced Cooling in
Fluid Nitrogen at High Densities and Temperatures," Phys. Rev,
Lett, (in press - November 10, 1986 issue).
6. R. Reichlin, 0. Schiferl, S. Martin, C. Vanderborgh, and R. L.
Mills, "Optical Studies of Nitrogen to 130 GPa," Phys. Rev, Lett.
55, 1464 (1985).
7. M. Ross and F. H. Ree, "Repulsive Forces of Simple Molecules and
Mixtures at High Density and Temperature," J. Chem. Phys. 73, 6146
(1986).
8. D. J. Stevenson, "Condensed Matter Physics of Planets: Puzzles,
Progress and Predictions," in High Pressure in Science and
Technology, Vol. Ill, edited by C. Homan, R. K, MacCrone, and E.
Whalley (North-Holland, New York, 1984), pp, 357-368,
143
GRAVITY DATA ANALYISIS
William L. Sjogren
Jet Propulsion Laboratory, Pasadena, CA 91109
liree different efforts kave been worked^ He first is the redue-
tion of raw Doppler data from the Apollo 15 snbsatallite to produce
acceleration profiles as a function of latitude, longitude and altitude
(~15,000 new observations). The second is an investigation related to
fitting long arcs of Pioneer Venus Orbiter tracking data. The third
effort was the study of gravity / topography ratios which were found to
have a linear trend with longitude.
Raw Doppler data from the Apollo 15 subsatellite have been archived
for the past 15 years. These data have now been reduced and put in a
form compactible with user-friendly software (Geophysical Data Facility
(GDF) which is being sponsored by the Lunar and Planetary Institute).
Approximately 150 orbits (~15,000 acceleration observations) covering a
+ 30° band of latitude with varing space altitude are plotted and des-
cribed in a report entitled "User's Guide to the JPL Doppler Gravity
Data Base" JPL publication 86-16. These data have been distributed to
NSSDC and LPI.
Fitting long arcs of Pioneer Venus Orbiter data (>3 orbits) became a
problem a few years ago when a global 10th degree and order gravity field
(Mottinger et al, 1985) was being estimated. Only realistic results for
the field could be obtained when short arcs of data were used (~1 to 2
orbits). Since then many tests and empirical models have been evaluated
to explain the poor fits to the observations. Presently we are obtaining
some encouraging results where excellent data fits have been obtained
over eight orbits. It is somewhat premature to state that the previous
problem is understood. It appears that significant perturbations are
being produced by the high frequency (local) gravity anomalies. The
introduction of many near surface mass disks in a global sense is very
effective. This work will continue.
The ratios of observed gravity to theoretical gravity based on
topography (p = 2.7 gm/cc) were calculated for 35 well defined Venusian
gravity anomalies as displayed by Sjogren, et al. 1980. The anomalies
were significantly greater than 5 milligals and were located between -15°
latitude and 42.5° latitude. Theoretical gravity profiles were generated
using USGS topography (Pettengill, et al., 1980) gridded on 2° s 2°
spacing. The extent of the topography was 60° east and 60° west of the
feature being evaluated. It also extended from 60°N to 30°S and was
isostatically compensated uniformly at 100 km. Some typical profiles for
four features are shown in figure 1. The dotted line is gravity while
the solid line is theoretical gravity based on topography having a
density of 2.7 gm/cc and being compensated at 100 km. Figure 2 shows the
plot of the ratios of gravity to topography. There appears to be an
increasing trend in the ratios from 60°E longitude. At least two of
these ratios had previously been confirmed by other investigations
(Bills, Esposito, Reasenberg) for Aphrodite Terra and Beta Regie.
Shallower compensation for Aphrodite Terra and deeper compensation for
Beta Regio are the generally accepted statements in the literature.
These new results seem to indicate that there is a definite trend east-
ward and possibly some internal dynamic process may be the cause of it.
Certainly the 300 km isostatic depth of compensation for Beta Regio
(Esposito, et al.) seems unrealistic and therefore dynamic forces sup-
144
porting it would be a more acceptable model. If tliis is so, tils result
may indicate a time sequence in the age of topography^ This report is
presently in peer review at Icarus where Sjogrens Trager and Saunders are
authors.
REFERS! CES
1. Bills* B.G. (1983), Gravity^ topography and orustal CYolutioa of ¥enus,
Icarus. 56 , 345-371.
2. Esposito, P.B,, W.L. Sjogren, N.A. Mottinger, B.G, Bills, and E. Abbott
(1982), ¥enus gravity: Analysis of Beta letio, Icarus , 51 448-459,
3. Mottinger, N,A., W.L, Sjogren and B.G. Bills (1985), Venus Gravity: A
Harmonic Analysis and Geopysical Implications, Jj. Geophys. Res..
90, C739-C756.
4. Pettesgill, G.G., E. Eliason, P.G. Ford, G.B. Loriot, 1. Masursky, and
G.E. McGill (1980), Pioneer Venus radar results: Altimetry and
surface properties, i. Geophys. Res., 85 , 8261-8270,
5. Sjogren W.L. S.J. Phillips, P.W. Birkeland, and 1,N. Wimberly (1980),
Gravity anomalies on Venus, J^. Geophvs. Res.. 85* 8295-8302.
6. Reasenberg, R.D., A,M. Goldberg, and I.I. Shapiro (1982), Venus: Compar-
ison of gravity and topography in the vicinity of Beta Regio,
Geophys. Res, Lett., 9, 637-640.
19.3
2.3
-a 3
72.1
20.2
-3L70
35.3
-2L5
-90
GUINEVERE
iLS
ORBIT 556
A'
33
NIOBE
k
/
N
'>.-■'
V
1
w
ORBIT 445
S.3
E. aphrodite/ A (^14
// \\G
14 AND M
-2.4
-3a 1
ORBIT 474
90
WTITUDE. deq
Figure 1
145
3.01 1 1 1 1 1 1 1 I I > I"
2.6
2.3
1.9
1.5
1.1
as
0.4
15
A
O 20 B
N 1« ,
17 ^
19
18
-^ 1-
>
a:
o
12
- 12
B /te
LETTERS REPRESENT NEGATIVE FEATURES
NUMBERS REPRESENT POSITIVE FEATURES
I I : I 1 L
'III L
60 120 180 240 30)
VENUS LONGITUDE (deg)
60
Figure 2
146
The Origin of Polarity Asjnnmetries in the History of the Geomagnetic
Field
E. H. Levy, Lunar and Planetary Laboratory and Department of Planetary
Sciences, University of Arizona, Tucson, Arizona 85721.
Paleomagnetic studies of the behavior of Earth's magnetic field
suggest that the field has exhibited persistent polarity asymmetries
throughout recent geologic time, with relative amplitudes of the lowest
several multipole moments differing between the "normal" polarity state
and the "reversed" polarity state. This paper examines the behavior of
magnetohydrodynamic stationary modes in the presence of an imposed weak
magnetic field originating separately from the dynamo. A rare class of
stationary states is found that exhibit high sensitivity to the
presence of weak imposed fields. The amplitude of the difference
between the total fields of opposite polarity is much larger than the
amplitude of the imposed nondynamo fields. It is proposed that Earth's
magnetic field operates in such a mode, highly sensitive to the
presence of an ambient field. An argument, based on the possible
mechanisms of dyanamical equilibration of dynamo magnetic fields, is
given to explain why the terrestrial dynamo should choose to operate in
one of these rare states. Implications are discussed for the general
mechanism of dynamo magnetic field equilibration in planets .
Magnetic Flares in the Protoplanetary Nebula and the Origin of
Meteorite Chondrules
E. H. Levy and S. Araki, Lunar and Planetary Laboratory and Department
of Planetary Sciences, University of Arizona, Tucson, Arizona 85721.
Meteoritic chondrules apparently resulted from very rapid,
transient, and short-lived heating events in the otherwise much cooler
protoplanetary nebula where the meteorites formed. This paper proposes
and analyzes a model for the chondrule forming heating events based on
magnetohydrodynamic flares in the corona of the protoplanetary nebula
which precipitate energy in the form of energetic plasma along magnetic
field lines down toward the face of the nebula. It is found that flare
energy release rates sufficient to melt prechondrular matter, leading
to the formation of chondrules , can occur in the tenuous corona of a
protostellar disk. Energy release rates sufficient to achieve melting
require that the ambient magnetic field strength be in the range that
has been inferred separately from independent meteorite remanent
magnetization studies.
147
Generation of a Dynamo Magnetic Field in a Protoplanetary Accretion
Disk T. Stepinski and E. H. Levy, Lunar and Planetary Laboratory and
Department of Planetary Sciences, University of Arizona, Tucson,
Arizona 85721.
A new computational technique is developed that allows realistic
calculations of dynamo magnetic field generation in disk geometries
corresponding to protoplanetary and protostellar accretion disks . The
approach Is of sufficient generality to allow, in the future, a wide
class of accretion disk problems to solved.
In this initial study, the basic modes of a disk dynamo are
calculated. Spatially localized oscillatory states are found to occur
in Keplerian disks . A physical Interpetatlon is given that argues that
spatially localized fields of the type found in these calculations
constitute the basic modes of a Keplerian disk dynamo.
These results and the computational technique have general
applicability to a variety of other cosmical disk systems including
disk galaxies and high energy accretion disks around black holes and
compact stars .
The Steady State Toroidal Magnetic Field at the Core-Mantle Boundary
S. J. Pearce and E. H. Levy, Lunar and Planetary Laboratory and
Department of Planetary Sciences, University of Arizona, Tucson,
Arizona 85721.
Recent measurements (Lanzerotti, et al., 1985) indicate that the
strength of the toroidal magnetic field at Earth's core-mantle boundary
is comparable to the strength of the poloidal field- -5 to 10 gauss.
Illustrative calculations are given to show that this result is an
inevitable result of the external boundary condition on the core, in
which the mantle electrical conductivity is several orders of magnitude
lower than that of the core . The measurements are shown to imply that
the internal core toroidal magnetic field is in the range of several
hundred gauss. Thus the measurements imply that Earth's core contains
a strong toroidal magnetic field and support the idea that Earth's
dynamo- -and, by implication, other planetary magnetic fields- -involves
efficient toroidal magnetic field generation through strong
differential rotation.
148
LUNAR MAGNETIZATION CONCENTRATIONS (MAGCONS) ANTIPODAL TO YOUNG LARGE IMPACT
BASINS, R. P. Lin, and K. A. Anderson,* Space Sciences Laboratory, University of California, Berkeley, CA
94720, L. L. Hood, Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ 85711
Electron reflection measurements from the Apollo 15 and 16 subsatellites show that patches of strong
surface magnetic fields ranging in size from less than ~7 km, the resolution limit of the observations, to greater
than 600 km, are distributed over the surface of the Moon (1, 2). With the exception of a few regions — Rima
Sirsalis (3) and Reiner Gamma (4) — no obvious association to surface geology has been found. In the ±35°
latitude band covered by the electron reflection measurements, the largest concentration of surface magnetic
fields extend in a chain from ~E160 S30 east and south to ~Wll5 S20 (Figure 1). In the north a large magcon
is located at ~Nl0-25 E85-110. The southern chain includes the intense magnetic region near Van de Graaff
initially found from Apollo subsatellite magnetometer measurements (5). The orbiting magnetometer also
observed the anomaly centered at ~W125 S23, at the northeastern end of the southern chain (6). We have pre-
viously noted that these large regions are approximately antipodal, i.e. located diametrically opposite, to large
circular impact basins (7). The southern chain appears to be composed of three regions corresponding to the
antipodes of the Imbrium, Serenitatis and Crisium basins. The large northern ma^etized region is antipodal
to Orientale.
We have examined the antipodes of the 23 ringed impact basins identified by Wilhelms (8) for which
electron reflection measurements are available. Measurements (in m° x W pixels) located within an antipodal
region equal in diameter to the basin itself were used to form distributions of the surface magnetic fields for
each basin (Figure 2). The median surface magnetic field for each basin antipode is plotted in Figure 3, where
the basins are arranged in order of increasing age. Strong magnetic fields are obtained for antipodes of the
young ringed impact basins, but significantly weaker fields are found for older basins. Strong median magnetic
fields are obtained at the antipodes of three young ringed impact basins, Orientale, Imbrium and Serenitatis.
The median magnetic field at the antipodes of 20 other ringed basins are significantly weaker. With one possi-
ble exception, all these basins are older than the three young basins just referred to.
Large impacts could produce significant modification of the antipodal region either by focusing of impact
generated seismic waves (9) or by clustering of secondary ejecta. Seismic waves may produce and enhance deep
fractures in the antipode region. If an underlying layer of the lunar crust was uniformly magnetized over a
very large region, possibly over the entire Moon, then the surface field from such a layer will be small except at
the edges of the region. Deep fractures produced at the antipodes of impact basins will result in discontinuities
in the magnetized layer, which would result in strong surface magnetic fields.
Alternatively, antipodal magnetic anomalies could mean that the basin-related secondary ejecta may
have become highly magnetized, possibly in a strong ambient or impact generated magnetic field.
The apparent temporal variation of the magnetic fields for the basin antipodes may reflect real variations
in the lunar magnetic field. Paleomagnetic data (10) suggest that a lunar "magnetic epoch" of strong fields
occurred approximately at the time of formation of the youngest impact basins. However, the lack of strong
antipodal fields for older basins may also be the result of more extensive gardening which has removed any
strong magnetization signature.
REFERENCES
(1) Lin R. P., Anderson K. A., Bush R., McGuire R. E., and McCoy J. E. (1976) Proc. Lunar Sci. Conf.
7th, pp. 2691-2704.
(2) Lin R. P. (1979) Proc. Lunar Sci. Conf. 10th, pp 2259-2264.
(3) Anderson K. A.', Lin R. P., McGuire R. E., McCoy J. E., Russell C. T., and Coleman P. J. Jr. (1977)
Earth Planet. Sci. Lett. 34, pp. 141-151.
(4) Hood L. L., Coleman P. J. Jr., and Wilhelms D. E. (1979) Proc. Lunar Sci. Conf. 10th, pp. 2235-2257.
(5) Russell C. T., Coleman P. J. Jr., Fleming B. K., Hilburn L., loannides G., Lichtenstein B. R., and
Schubert, G. (1975) Proc. Lunar Sci. Conf 6th, pp. 2955-2979.
(6) Hood L. L. (1979) Papers presented to the Conference on Origins of Planetary Magnetism, 8-11
November 1978, Lunar and Planetary Institute, Houston, Texas.
(7) Lin R. P., El-Baz F., Hood L. L., Runcorn S. K., Schultz P. H. (1980) (abstract) Lunar Planet. Sci.
XI, p. 626.
(8) Wilhelms, D. E. (1979) NASA Rep. Planetary Geology Program 1978-1979 Washington, D.C., Tech.
Mem. 80339.
(9) Schultz P. H., and Gault D. E. (1975) The Moon 12, pp. 159-177.
(10) Cisowski S. M., CoUinson D. W., Runcorn S. K., Stephenson A., and Fuller M. (1983) / Geophys.
Res. 88, Suppl., pp. A691-A704.
* Also Physics Department.
149
SlWaCE MAG^CTK: FELD
OHSnL
► "•aanT
Q
o° o
7
'J>
o
if^aSet
iSSCMTATB
llt»Uri
«TAT)S ,^/N ,^~>.
o
3x10'^ cr
= 2x10^
E
O
tr.
UJ
Q.
1 J -! 1
LUNAR BASIN ANTIPODES
j-NECTARlS—
] rlMBRlUM
tq!.nn_
Figure 1 (above). A contour map of the strongest sur-
face magnetic fields measured by the electron
reflection method. The circles indicate the antipodes
of large impact basins.
12 3 4
MAGNETIC FIELD (nT)
Figure 2 (left). The distribution of W x 1H° mag-
netic field samples in the antipodal region for the Nec-
taris and Imbrium basins. The arrows indicate the
median field value.
11
» I »
s
so -r 00 M
»0 O O M
« CO (O «>
'''^»'-^ l-< "*— .
r-< «D --^ ■^r
eo ct 00 »<•
■^ O —»
00 O O
O OB CO
^ wo
•H i
o J
-NVIUUMI-
a .B a
E S ^
o. *i - —
e s " s
.« « *- =^
KcnO a
a ? -5
-NVlUViOaN-
3 „
I'll
T3 ti: .-
=5 a
a.-c
3
-r
M
■T
*o \n -f to
oo
CO OO CO o>
TT
eo
C4 C4 o eo
a
■bO
d
■§
a
2-s
>
aosov-F
m)
ditatis)
uilUtat;
1?
t
-S
'5
Lomo
Nubiu
Fecun
Traaq
Si?
a ^
NVI71VXC
bO
a
^.i.7
>?tN
■aar"^
Figure 3. The median magnetic field for the antipodal regions of 23 impact basins. The number of samples
and the total number possible for the antipodal region are given to provide an indication of the coverage.
150
Implications of Convection in the Moon and the Terrestrial Planets
Donald L. Turcotte, Department of Geological Sciences, Cornell University,
Ithaca, NY 14853-1504
During the past year we have worked on two principle lines of research:
1) The early thermal and chemical evolution of the moon. The work on the
evolution of isotopic systems on the moon has now been published (1) . The
rubidixom- strontium, neodymium- samarium, and uranixun- thorium- lead systems were
studied. The relation of source region heterogeneity to the mixing associated
with mantle convection was considered. We are continuing this line of
research by studying the chemical evolution of the moon and terrestrial
planets by means of parameterized convection. One aspect of our studies is
the concentration of trace elements including the heat producing elements into
the crust thus influencing the rate of mantle convection. A second aspect of
our studies is the delamination of planetary lithospheres as a cooling
mechanism and as a means of returning trace elements to the planetary inter-
ior.
2) Fractals. Our group is the only group working on the application of
fractal concepts to planetary geology and geophysics. With the current
explosion of interest in and research on fractals this promises to be a most
rewarding line of research. Our initial results have been quite exciting.
Our first application of the fractal concept was to fragmentation including
the frequency- size distribution of meteorites, asteroids, and particulate
matter produced by impacts. This work has now been published (2). We have
continued our fractal research on geophysical spectra. Global spectra are
available for topography and geoid on the Earth, Venus, Mars, and the moon.
If the spectral energy density has a power law dependence on wave number a
fractal is defined. The topography spectra for the earth is a well-defined
fractal with D = 1.5; this corresponds to Brown noise with the amplitude
proportional to the wave length. Although there is more scatter for the other
planetary bodies, the data for Mars and the moon correlate well with the data
for the earth. Venus also exhibits a Brown noise behavior but with a smaller
amplitude. The power law dependence of the Earth's geoid is known as Kaula's
law. We show that uncompensated Brownian topography gives a power law
dependence that is in good agreement with Kaula's law. However, the required
amplitude is only 8% of the observed topography. This ratio increases to 72%
for the moon. This work was presented at the 17th Lunar and Planetary Science
Conference, Houston, Texas, March 17-21, 1986 (3) and has been written up and
accepted for publication in the Proceedings (4) .
(1) D.L. Turcotte and L.H. Kellogg, Implications of isotope data for the
origin of the Moon, in Origin of the Moon . W.K. Hartmann, R.J. Phillips and
G.L. Taylor, eds . , Lunar and Planetary Institute, 311-330 (1986).
(2) D.L. Turcotte, Fractals and fragmentation, J. Geophys. Res., 9J., 1921-
1926 (1986).
(3) D.L. Turcotte, A fractal approach to planetary spectra, Lunar Planet.
Sci. XVII, 905-906 (1986).
(4) D.L. Turcotte, A fractal interpretation of topography and geoid spectra
on the earth, moon, Venus, and Mars, J. Geophys. Res., in press (1986).
151
Global Petrologlc Variations on the ^foon: A Ternary-Diagram Approach
Philip A« Davis and Paul D„ Spudis^ Branch of Astrogeology, U,S« Geological
Surveys Flagstaff s Arizona 86001
Approach^ We have used the ternary-diagram approach outlined herein in an
attempt to show on a single map as much detailed geochemical information
indicating petrologic variations within the lunar crust as possible. We
confine the presentation of our ternary-diagram analysis and subsequent
discussion of the results to use of the Fe (wt%) versus (Th/Ti) variation
diagram, because these data include only the Apollo gamma-ray orbital data,
which have more global coverage (about 19%) than that of the X-ray orbital
data (about 9%). The (Th/Ti) ratio represents the observed Th/Ti ratio
normalized to CI chondrite values » We produced an error database for Fe and
(Th/Ti) by the standard method of determining error from counting statistics
to take into account the analytical errors associated with the Fe and (Th/Ti)
data in the present analysis.
The method starts by establishing a ternary reference diagram whose three
sides are each divided into eight segments. Each ternary subdivision is
assigned a distinct color; the colors represent a spectral continuum from red
to green to blue to red* The center of the diagram is a trianglar area
representing approximately equal proportions of each apex, and it is thus
assigned a gray color. Assignment of rock end-member compositions to the
three apexes allows rock or soil compositions that are binary or ternary
mixtures of these three end members to be represented as continuous colors in
the visible spectrum^
Each of the end-member Fe and (Th/Ti) compositions is an average
calculated from values reported in the literature. The W^-suite (troctolite
and norite) and KREEP rocks are represented by the red apex, the mare basalt
by green, and the ferroan anorthosite by blue. Thus, we can now assign a
color from the ternary reference diagram to each pixel in the orbital
geochemical databases, using the Fe~concent ration and (Th/Ti) -ratio values of
the pixel and the ternary apexes. The relative proportion of each of the
three compositional end members (apexes) needed to produce the observed
composition of a particular pixel is determined trigonometrically. Once the
pixel's location within the diagram has been determined, the color at that
location is assigned to that pixel's position in a new image or map file.
This process is then repeated for every pixel within the orbital databases.
Also, the frequencies of occurrence of pixels at a particular ternary
composition are accumulated within a ternary scattergram.
Certain pixels within the Fe and (Th/Ti) databases have high
uncertainties^mostly because of their low orbital accumulation times. To
determine the effects of these errors on the areal abundance of units within
the classification map, we decreased by increments the amount of error that
pixels could have in terms of Fe and (Th/Tl) before being excluded from the
classification map. The following discussion pertains to the classification
map (not shown) in which pixels that have errors greater than 75% are
excluded. This level of error exclusion provides a reasonable amount of
certainty for the remaining unit pixels and their areal abundances,
Mscussioa, Examination of the classification map allows easy determination
of (1) the global spatial distribution of end-member compositions, (2) the
transitional spatial relations between end-member compositions, and (3)
quantitative estimates of the relative proportions of each end member at each
pixel location within the orbital groundtracks. The use of elemental ratios
in our analyses , instead of the commonly used elemental bivariate diagrams
152
[1^2], shows geologic information that is otherwise hidden in individual
elemental databases «
The Apennine Bench region is shown to have a composition corresponding to
a mixture of KREEP and mare basalt 5 which is consistent with the results of
previous studies [3,4]. Other areas of t^-suite/KREEP material are in the
farside highlands near Van de Graaff (18°-29°S, 175°-171°W), within the
Hertzsprung basin (3°-5°S, IZS'-ISO^W), and south of ffere Smythii and west of
Pasteur Crater (7°-15°S, 76°-98°E). The first two of these KREEP-rich
highland areas coincide with areas of highland crustal thinning [5^6] that are
covered by the orbital gamma-ray data. An inverse relation between Th
concentration and highland crustal thickness has been reported by [7]. The
preliminary elemental concentrations obtained by [8] suggest that the Van de
Graaff region may have a "granitic" rock composition, similar to that of
sample 12013* Generally, the average composition of lunar granites is lower
in Fe, Ti, and liig and significantly higher in K and Th than that of KREEP
basalts. The Fe and Ti concentrations of these three highland areas are
indeed lower than that at any of the three nearside high-KREEP areas, possibly
because of the proximity of maria to the KREEP-rich areas; however, the three
highland KREEPy areas do not appear to be associated with extensive mare
deposits* For K, [9] have presented preliminary gamma-ray data showing the
Van de Graaff region to have only 880 ppm K, whereas they report the Fra Mauro
region as having 2680 ppm K. This significantly lower K value for Van de
Graaff strongly suggests that these KREEP/Wig-suite highland areas (at least
those near Van de Graaff) are not composed of "granitic" rock. They are most
likely either "KREEPy basalts" [10] resulting from volcanism propagated by
crustal thinning in these areas, or the remnants of an 1%-suite pluton exposed
by an early impact event (such as the South Pole-Aitken basin; [11]).
Our classification map also shows that, at the spatial resolution (about
100 km) of the gamma-ray instrument, the central regions of most major maria
have relatively pure mare-basalt compositions. Only Mare Tranquillitatis
appears to have compositions transitional between mare basalt and ferroan
anorthosite, which is probably the result of the addition of underlying
anorthositic highlands debris to mare-basalt regoliths by vertical mixing
through relatively thin, young, blue mare-basalt flows [12,13]. At
Aristarchus, the unit map indicates a mixture of KREEP, mare basalt, and
ferroan anorthosite, which grades into a more KREEP- and mare-basalt-rich unit
at the north border of the groundtrack. The presence of these two units can
be attributed to the relatively thin, young, blue mare-basalt flows in these
two areas [14] that mixed with underlying KREEP- or ffe-suite-rich highland
terrain. This underlying material is present at relatively shallow depth, as
indicated by its exposure within Aristarchus Crater [15,16].
A series of relatively young lava flows with well-developed flow fronts
occur in southwestern Jfere Imbrium [17], In addition to their striking
morphological development, remote-sensing data indicate that these lava flows
are rich in Ti [14], relatively rich in Th (8,0 ppm; [18]), and young (less
than 2.0 b.y. old; [19]). These high-Ti, KREEP-rich lavas are represented on
our petrologic map by two units of mostly mare basalt with some KREEP
component .
Another interesting area is within and near the Balmer basin on the lunar
eastern limb (lO^-lS^S, 75°E). The position of this zone correlates with the
light plains fill of the Balmer basin, which has been described previously as
KREEP-rich, mare-like deposits [10,20]. Our classification map shows these
plains to represent roughly an equal mixture of anorthosite, mare basalt, and
KREEP/Mg-suite material. The identification of dark-halo craters in this
153
region [20,21] supports the suggestion that light plains in the region thinly
mantle buried, KREEP-rich mare-basalt flows. These basalt flows are probably
older than 3.9 b.y. because they are buried by highland plains of Imbrian to
Nectarian age [22]. It thus appears that the ancient lunar maria (older than
3.9 b.y.) had a diversity of chemical compositions, ranging from "normal"
chondritic Th/Ti values to more KREEP-rich varieties.
The lunar surface represented by the Apollo orbital groundtracks is shown
to consist of 8.4% relatively pure (85%) ferroan anorthosite, even though
pixels that have high compositional uncertainties were excluded. Deleting the
maria from these data raises this value to 12.9%. Most of the lunar highlands
is composed of four units. Considering the areal percentages of the
groundtracks and the modal amounts of the end-member components for these four
units results in an average highland composition of 68% ferroan anorthosite,
29% mare basalt, and 3% KREEP/Mg-suite rocks. This resultant rock composition
approximates that of "anorthositic gabbro" and is consistent with our previous
analyses . [23] . This composition may represent the average composition of the
upper half of the highlands crust [24] .
Significant amounts of mare basalt (21,1% of the Apollo groundtrack)
occur within the highlands (mostly on the eastern limb and farslde highlands) ,
as indicated by the areal distribution of a unit composed of 65% ferroan
anorthosite, 3% KREEP/Mg-sulte, and 32% mare basalt. Its areal distribution
coincides with mapped occurrences of highland plains that display dark-halo
craters [19,21], for which spectral data Indicate the presence of excavated
mare basalt [25]. This coincidence suggests that mare volcanlsm occurred
within these highland areas before the end of the final heavy bombardment. We
do not, however, dismiss the posslbj.lity that this unit may represent some
type of highland gabbro or a mixture of Mg-suite rocks with an as-yet-
unsampled mafic rock type that has a subchondritlc Th/Ti ratio.
References. [1] Clark, P. E. et al. (1978) PLPSC 9 , 3015; [2] Clark, P. E.,
and B. R. Hawke (1981) PLPS 12B , 727; [3] Hawke, B. R. , and J. W. Head (1978)
PLPSC 9 , 3285; [4] Spudis, P. D. (1978) PLPSC 9 , 3379; [5] Frontispiece (1976)
PLSC 7 ; [6] Bills, B. G., and A. J. Ferrari (1977) JGR 82 , 1306; [7] Metzger,
A. E, et al. (1977) PLSC 8 , 949; [8] Metzger, A. E. et al. (1974) PLSC 5 ,
1067; [9] Parker, R. E. et al. (1981) LPS XII , 811; [10] Hawke, B. R., and
P. D. Spudis (1980) Proc Conference on the Lunar Highlands Crust , 467;
[11] Wilhelms, D. E. (1984) Moon, in The Geology of the Terrestrial Planets ,
NASA SP-469, 106; [12] Horz, F. (1978) PLPSC 9 , 3311; [13] Whltford-Stark,
J. F., and J. W. Head (1980) JGR 85 , 6579; [14] Pieters, C. (1978) PLPSC 9 ,
2825; [15] McCord, T. B. et al. (1972) JGR 77 , 1349; [16] Guest, J. E., and
P. D. Spudis (1985) Geol. Mag. 122 , 317; [17] Schaber, G. G. (1973) PLSC 4 ,
72; [18] Etchegaray-Ramlrez, M. I. et al. (1983) PLPSC 13 , JGR 88 , A529;
[19] Schultz, P. H. and P. D. Spudis (1983) Nature 302 , 233; [20] Hawke, B. R.
et al. (1985) Earth Moon Planets 32 , 257; [21] Schultz, P. H., and P. D.
Spudis (1979) PLPSC 10 , 2899; [22] Wilhelms, D, E., and F. El-Baz (1977) USGS
Map 1-948 ; [23] Davis, P. A. and P. D. Spudis (1985) PLPSC 16 , JGR 90 , D61;
[24] Spudis, P. D., and P. A. Davis, In press, PLPSC 17 , JGR; [25] Hawke,
B. R., and J. F. Bell (1981) PLPS 12B, 665.
154
A CHEMICAL AND PETROLOGICAL MODEL OF THE LUNAR CRUST.
Paul D, Spudis and Philip A. Davis, U.S» Geological Survey^ Flagstaff j, AZ
86001
It has long been recognized that the lunar crust is chemically and
petrol ogically heterogeneous j although the exact details of its structure
remain contentious. Various models of lunar crustal structure have been
proposed that invoke lateral heterogeneity [e.g., l]j vertical variations
'2s3]5 or both [4]» Recently, we presented the results of our attempts to
make petrol ogic maps of the lunar highlands from orbital geochemical data
[5]. These maps provide data on petrologic variations in the uppermost
lunar crust. Moreover, recent studies have clarified the process involved
in multi-ring basin formation [e.g., 6, 7], resulting in models that
permit us to use basin as probes to understand the composition of the
lower crust. Thus, it is now appropriate to reevaluate a variety of
remote-sensing and lunar sample data to determine the probable chemical
and petrologic structure of the lunar crust.
Composition and structure of the upper crust . The average composition of
the highland crust under the Apollo 15 and 16 orbital groundtracks is that
of "anorthositic gabbro" (AI2O3 26-28%) [5]. Within this area of coverage
is considerable petrologic variation, ranging from deposits of "pure"
anorthosite to KREEP basalt [5]. Our efforts to map highland rock types
used the Fe - Th/Ti chemical plot; such a variation diagram readily
separates the ferroan anorthosites (believed to be original crustal
products of lunar primordial differentiation [8]) from the Mg-suite rocks
and KREEP. Of the eleven petrologic units mapped in [5], five have
"chondritic" Th/Ti ratios, indicating affinity with the ferroan
anorthosite suite rather than with the Mg-suite or KREEP. These units
make up over 80% of the toal highland region covered by the Apollo orbital
data. Assuming that the Apollo groundtracks are representative of the
whole Moon, we conclude that rocks of the Mg-suite represent a minor
fraction of the upper lunar crust. The dominance of Mg-suite rocks in the
sample collection is probably more apparent than real and may be mostly
due to the emphasis placed on study of pristine clasts from highland
breccias.
A consequence of the heavy bombardment of the lunar crust was the
developnent of a thick, brecciated debris layer, the "megaregolith" [9].
The thickness of this layer has been estimated as 1-2 km [10-11] to as
much as 40 km [12], Because of the general paucity of pristine plutonic
rocks and the lunar cratering history implied by the apparent 3.92 b.y.
age of the Nectaris Basin [13,14], we favor values of tens of kilometers
for megaregolith thickness. Mixing in such a debris layer would be
dominantly vertical, analogous to mixing in the thinner regoliths
developed on mare basalt flows. Such mixing would preserve "outcrops" of
pure rock types such as those on the petrologic maps [5].
Composition of the lower lunar crust . Evidence for lower crustal
composition comes from the composition of multi-ring basin ejecta. Of the
eleven basins covered by Apollo orbital chemical data, nine display
enrichments of norite in their continuous near-rim deposits, in contrast
155
to the more anorthositic compositions of the interbasin terrain.
Moreovefj the relative fraction of norite in basin ejecta determined by
mixing models increases with increasing basin diameter [4^5]. Assuming
these relations to be non-coincidental ^ we suggest that the lower crust is
composed dominantly of noritic rocks (AI2O3 -u 20%). Additional evidence
for this noritic composition comes from the (probably) polymict "Very High
Aluminum "(VHA) and "low-K Fra Mauro" (LKFM) basalt rocks, which appear to
be basin impact melts [14]; these rocks are noritic in bulk chemistry.
Given estimated formation energies of impact-basins [15,16] and projectile
penetration depths [6], we consider most basin impact melts to have been
generated at depths of about 30 to 60 km in the lunar crusts If VHA and
LKFM basalt does represent basin impact melt, these rocks provide direct
evidence for a noritic lower crust. To explain the lack of lunar mantle
material in impact melts, Ryder and Wood [2] suggested that a noritic LKFM
layer exists midway in the crust. Analysis of both the geologic settings
of the Apollo landing sites [17,18] and particle movements during
cratering flow [6,19] suggest such a crustal structure is not a
requirement, as the basin impact melt sampled during Apollo missions
should be contaminated with only shallow-level clasts. We believe that
the available data are consistent with the hypothesis that the lower lunar
crust is composed dominantly of norite,
A lunar crustal model . Our lunar crustal model is shown in Figure 1. It
has essentially two-layers: anorthositic mixed rocks overlie a generally
noritic crystalline basement. The contact between these layers is
probably gradational on a scale of kilometers. The Mg~suite comprises a
series of plutons, some of them layered, that are a subordinate fraction
of the crustal volume. In this model, the crust is laterally and
vertically heterogeneous on a scale of tens of kilometers.
Our model has several implications for lunar crustal origin and
evolution. Concurring with previous suggestions (see review in [8]), we
believe that a global magma system produced an original ferroan
anorthosite crust by plagioclase flotation and oli vine/pyroxene sinking.
As the crust grew downward and the mantle upward, the residual liquid had
a noritic composition [2,20]; this layer formed the noritic lower crust
and eventually, a layer of material (KREEP substratum; Fig, 1) greatly
enriched in incompatible trace elements [20,21]. Partitioning of
incompatible elements into this KREEP layer was not totally efficient,
resulting in the production of KREEP-rich norites in the lower crust
[2]. Concurrent with this episode was intrusion of Mg-suite rocks to form
plutons within the original anorthositic crust. This activity was
virtually complete by 4,36 b,y, ago [22]. Massive extrusions of mare
basalts [5] and continued heavy bombardment of the crust followed until
about 3.85 b.y. ago. During this interval, the hot, low-density KREEP
material migrated upward [23] and was disseminated through most of the
lower crust. Thus, basins forming around 3,9 b.y. ago incorporated a
KREEP trace-element pattern into their noritic impact melts ("Melt rocks";
Fig. 1), In the Imbrium region, KREEP basalts were erupted on the surface
both before [24] and after [25] the Imbrium impact. After the heavy
bombardment ended, mare basalts continued to be extruded, forming the
visible maria. Continued study of lunar samples and new remote- sensing
156
data from the proposed Lunar Geoscience Observer Mission will be required
to assess the validity of this lunar crustal model.
References: [1] James, O.B. (1980) PLPSC H, 365. [2] Ryder, G. and
Wood, J. A, (1977) PLSC 8., 655. [3] Charette, M,P. et al . (1977) PLSC 1,
1049. [4] Spudis, P.O. et al . (1984), LPS XV, 812. [5] Davis, P,A, and
Spudis, P.0» (1985) PLPSCli, JGR 90, D61, T6] Croft, S.K. (1981) PLPSC
12A, 207, [7] Spudis, P.O. et aUTl984) PLPSC 15, JGR 89, C197» [8]
Warren, P,H, (1985) Ann, Rev, Earth Planet. Sci, J^, 201. [9] Hartmann,
W,K, (1973) Icarus 18, 634. [10] Short N. and Forman, M, (1972) Mod, Geol .
3, 69, [11] Horz, F. et al. (1976) PLSC ^s 2931, [12] Cashore, J. and
Woronow, A, (1985) PLPSC 15, JGR 90, C811, [13] James, O.B. (1981) PLPSC
12B , 209, [14] Spudis, P,D. (1984T PLPSC 15, JGR 89 , C95. [15] Baldwin,
R. (1963) Measure of the Moon , U, Chicago, 488 p, [16] Grieve, R.A.F. and
Dence, M,R, (1979) Icarus 38, 230, [17] Head, J.W. and Settle, M, (1976)
Imb, Consort, 1, 5, [18] Spudis, P,D, (1982) Ph,D, Thesis, ASU, 291 p,
[19] Grieve, R,A,F, et al. (1977) Imp, Explos, Crater , 791, [20] Wood,
J. A. (1972) Icarus 16, 462, [21] Warren, P,H, and Wasson, J.T, (1979)
R6SP 17, 73. [22] Lugmair, G,W, and Carlson, R.W. (1978) PLPSC 9., 689,
[23] Shirley, D,N. and Wasson, J.T, (1981) PLPSC 12B, 965, [24] Hawke,
B.R. and Head, J.W, (1978) PLPSC 9, 3285. [25] Spudis, P,D. (1978) PLPSC
1, 3379,
Bulk Compositions
Former shallow layered pluton
Vertical
Gabbroic anoithosite scale
Whole Melt Crustal '*'lf^f;„^»"'y Ma-anortho»it«. Mare basalt .north, norlte
crust _rocks. lasers ,„^'",5°„, ^
Nomenclature
AI2O3
1.
"Anorlhositic norite'
24-25%
2.
■VHA basalt"
20-25%
3.
"LKFM basalt"
18-22X
4.
"Anorlhositic gabbro"
26-28X
5.
"Norite"
Z0%
Megaregolith
KREEP substratum
4©
80km
Anorth. norite
norite
KREEP-norite
Figure 1, Proposed model of the lunar crust, FAN refers to ferroan
anorthosite. Provenance of bulk compositions indicated at left by
arrows. See text for discussion.
157
CALCIUM CARBONATE AND CALCIUM SULFATE IN MARTIAN METEORITE EETA79001.
J. L. Gooding (a) and S. J. Wentworth (b) . (a) SN2/Planetary Materials
Branch, NASA/Johnson Space Center, Houston, TX 77058. (b) Lockheed/EMSCO,
Mail Code C23, NASA/JSC, Houston, TX.
Introduction . If shergottite meteorites are rocks ejected from Mars by-
impact cratering, then they might be expected to contain samples of Martian
weathering products. Indeed, sulfur- and chlorine-rich minerals of
apparently pre- terrestrial, low- temperature origin were previously found in
samples of the shergottite. Elephant Moraine, Antarctica A79001 (EETA79001)
[1,2]. Further work on additional samples has revealed the occurrence of
calcium carbonate [3] and calcium sulfate in the same meteorite.
Occurrence of Carbonate and Sulfate . Chips of glassy Lithology C of
EETA79001 were studied by scanning electron microscopy (SEM) and energy-
dispersive x-ray spectrometry (EDS) to determine the mineralogy and
petrogenesis of the glass that was shown by others [4,5] to contain trapped
Mars -like gases. Calcium carbonate was identified as massive to acicular
crystals for which Ca, C, and were the major elements. Calcium sulfate
was identified as prismatic -acicular crystals with Ca and S as the major
elements . Additional SEM/EDS work is in progress to better more identify
the minerals and their parageneses .
Ca-carbonate occurs in at least three locations in EETA79001, all of
which are closely associated with glassy Lithology C. The largest deposit
occurs as a drusy halo around a large glassy vug at the center of the
meteorite. A second, more modest deposit occurs near the large deposit but
in association with a separate, smaller glass pocket. The third deposit
was recognized only after SEM/EDS reconnaissance of a chip from its
glass -pocket host. The two macroscopically visible central deposits are
not connected to the surface of the meteorite by any obvious system of
fractures. In addition, parallel studies of demonstrably Antarctic
(surface -located) weathering products in EETA79001 and other meteorites
from the same field locality [6] failed to find Ca-carbonate as a
terrestrially formed phase. Therefore, present physical evidence supports
pre- terrestrial origin of the carbonate. Stable- isotopic analyses of C and
in the carbonate are in progress by our collaborators and should help to
positively establish whether the carbonate formed in Antarctica. If an
Antarctic origin can be excluded, then origin of the carbonate on Mars
would become the favored interpretation.
Pre-terrestrial origin of Ca-sulfate in EETA79001 is less clear.
Gypsum is a known Antarctic weathering product in the exterior of EETA79001
and in other Antarctic meteorites [6], However, the sole occurrence of
Ca-sulfate documented to date in interior samples of EETA79001 is in the
intensively studied sample 27 glass pocket which has yielded the highest
concentrations of trapped Mars-like gases [4,5]. Furthermore, the possibly
pre-terrestrial Ca-sulfate occurs as euhedral crystals, some of which are
crystallographically oriented inclusions in quench- textured pyroxenes.
Other Ca-sulfate crystals in the same sample are intimately associated with
Ca-carbonate and also appear to be unmelted relics. The Ca-sulfate is
volxometrically less abundant than the Ca-carbonate and probably will not be
separable in sufficient quantities for conventional isotopic analyses.
Implications for Martian Geology . Gooding and Muenow [2] previously
hypothesized that origin of the unusual glass pockets in EETA79001, and
concomitant trapping of Martian atmospheric gases , was related to
shock- induced melting of pre-existing weathered (or deuterically altered)
areas in the target rock. New evidence for relict grains of carbonate and
158
sulfate minerals in the gas -rich glass supports the Gooding-Muenow
hypothesis, reiterating the unique nature of EETA79001 and the nearly
inescapable conclusion that it is a Martian rock. Accordingly, the case
for using EETA79001 and other shergottites for deriving geochemical
properties of Mars seems to be strengthened.
Various workers have speculated about possible storage of major amounts
of carbon dioxide as carbonates in the Martian regolith [7-9] although
attempts to find direct evidence for carbonates in remotely- sensed data for
Mars have produced negative results [10,11]. Therefore, if further work
demonstrates that Ca-carbonate in EETA79001 is pre -terrestrial in origin,
then we will have probably achieved the first confirmation of carbonates on
Mars. In that case, geological and paleoclimatological studies of Mars
need not dwell on whether carbonates exist but can concentrate on
determining locations, ages, and abundances of carbonate deposits.
References :
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Gooding J . L
Program - 1985
Gooding J . L
1049-1059.
Wentworth S
Bogard D. D
Becker R.
225-242.
Gooding J
Fanale F.
and D. W. Muenow (1985) Repts . Planet . Geol. Geophys .
NASA Tech. Memo. 88383, 161-163.
and D. W. Muenow (1986) Geochim. Cosmochim. Acta . 50 .
J. and J. L. Gooding (1986) Meteoritics . 21 . in press,
and P. Johnson (1983) Science . 221, 651-654.
H. and R. 0. Pepin (1984) Earth Planet . Sci. Lett. . 69,
L. (1986) Geochim. Cosmochim. Acta . 50 .
P. (1976) Icarus . 28, 179-202.
Kahn R. (1985) Icarus . 62, 175-190.
Haberle R. M. (1985) Nature, 318, 599-600.
Singer R. , T. B. McCord, R. N. Clark, J. B. Adams
(1979) J^ Geophys . Res. . 84, 8415-8426.
Roush T. L. , D. Blaney, T. B. McCord, and R. B.
Planet . Sci. XVII . Lunar and Planetary Institute
2215-2223.
and R. L. Huguenin
Singer (1986) Lunar
Houston, 732-733.
159
The Case for A Wet, Warm Climate on Early Mars
J.B. Pollack, J.F. Kasting
NASA Ames Research Center, Moffett Field, CA 9^035
S»M. Richardson
Iowa State University, Ames, Iowa 50011
K. Poliakoff
I.M.I. Inc., San Jose, CA
The theoretical arguments are presented in support of the idea that Mars
possessed a dense CO2 atmosphere and a wet, warm climate early in its
history. Calculations with a 1-D radiative-convective climate model
indicate that CO2 pressures between 1 and 5 bars would have been
required to keep the surface temperature above the freezing point of
water early in the planet's history. The higher value corresponds to
globally and orbitally-averaged conditions and a 305^ reduction in solar
luminosity; the lower value corresponds to conditions at the equator
during perihelion at times of high orbital eccentricity and the same
reduced solar luminosity.
The plausibility of such a CO2 greenhouse is tested by formulating a
simple model of the CO2 geochemical cycle on early Mars. By
appropriately scaling the rate of silicate weathering on present Earth,
we estimate a weathering time constant of the order of several times 10'
years for early Mars. Thus, a dense atmosphere could have persisted for
a geologically significant time period (~ 10^ years) only if atmospheric
CO2 was being continuously resupplied. The most likely mechanism by
which this might have been accomplished is the thermal decomposition of
carbonate rocks induced directly and indirectly (through burial) by
intense, global scale volcanism. For plausible values of the early heat
flux, the recycling time constant is also of the order of several times
10' years. The amount of CO2 dissolved in standing bodies of water was
probably small; thus, the total surficial CO2 inventory required to
maintain these conditions was approximately 2 to 1 bars. The amount of
CO2 in Mars' atmosphere would eventually have dwindled, and the climate
cooled, as the planet's internal heat engine ran down. A test for this
theory will be provided by spectroscopic searches for carbonates in
Mars' crust.
160
THE 1»LE OF REGOUTH Al^)RPnO^ M THE TRAMSmOM FmM
EffiLY TO MTE MmS CUMATE
F.P. Fanale, S.E. Postawko, A.P. Zent.and J.R. Salvail, Planetary Geosciences Div., Hawaii
Instit. of Geophysics, Univ. of Hawaii, Honolulu, HI 96822
The total degassed COg inventory on Mars is ciirrently thought to be the equivalent
of about 3.0 bars or less (l). Using a surface pressure of ~1 bar, several radiative
tremsfer models for early Mars have been advanced (2) proporting to show that a
significant greenhouse effect could have characterized the early Mars environment,
possibly explaining the intense channeling of the most ancient Mars terrain.
We have reexamined those models with regard to the effect that regolith adsorp-
tion may have had. In our model, we take into account: 1) the atmospheric greenhouse
effect, 2) the existence, mass, and temperature of any cap, 3) the partition of COg
between the atmosphere-cap system and the regolith as required by the latitudinal
temperature distribution, and 4) atmospheric heat transport.
We solve simultaneously for all these interdependent variables for cases involving
realistic total COg inventories, a variety of published greenhouse models, and both
current and lower solar constants.
To describe regolith adsorption properties we use the mineralogically insensitive
adsorption relationship developed by Zent et al. (3), which is normalized to be con-
sistent with observed surface properties of Viking-sampled soil (4). We should point out
that adsorption is relatively more effective under current Martian conditions than at
the higher temperature and pressure conditions that may have characterized the early
Martian environment. However, the regolith currently has adsorbed 10 - 100 times the
atmospheric COg inventory, whereas in most models the removal of only half the origi-
nal atmospheric COg inventory would be sufficient to destroy the strong greenhouse
effect.
Resulting scenarios fail into several classes. Either: 1) The greenhouse effect is so
weak that a cap exists throughout the period of growth of several hundred meters of
regolith. In this case the surface environment is essentially unaffected by the develop-
ment of any reasonably postulated regolith; or 2) A cap exists initially, but is
sufficiently small that the growth of >100m or so of regoliih causes a significant repar-
titioning of COg molecules among the three parts of the system and the equilibrium
solution finally involves no cap. In this case the regolith has no effect on the atmos-
pheric pressure and resulting temperatures until the cap disappears. However, once
the cap disappears, the effect of further regolith growth is to dramatically lower
atmospheric pressure and hence surface tem.peratures; or 3) A substantial greenhouse
effect exists initially, and there is no cap initially. In this case, also, regolith growth to
a depth of several hundred meters dramatically lowers atmospheric pressures and sur-
face temperatures, and the effect of the regolith is important from the outset.
In all scenarios which start out with surface pressures on the order of 1 bar, high
pressures and temperature conditions are truncated by the growth of only a few hun-
dred meters of regolith, ultimately leading to present Mars conditions. It might be
argued that while the development of a regolith a few hundred meters thick in early
Mars history is likely in view of the existence of the ancient cratered terrain, it is
vmreasonable to postulate a much lower regolith thickness at the very outset. We point
out, however, that it is not literally the thickness of the regolith that is the critical
parameter, but rather its total surfac area. Weathering products such as palagonites
and nontronites typically exhibit specific surface areas for COg adsorption of tens to
161
hundreds of square meters per gram, whereas basalts typically exhibit specific surface
areas of a fraction of a square meter per gram to several square meters per grami.
Thus, although we represent the "regolith growth" as a physical thickening (keeping
the specifc surface area equal to the Viking derived value), the acutal case involves the
product of both actual thickening and a great increase due to an increased abundance
of weathering products. While the precipitation of carbonates has probably been an
important process during Mars history (5). the rates at which this process could have
taken place under early Mars conditionswould have dropped sharply once liquid water
was fairly scarce. Furthermore, conditions under which liquid water was available may
have involved efTicient recycling of carbonate so that steady state conditions rather
than irreversible COg removal prevailed. In contrast, the growth of regolith surface
area demands corresponding and predictable COg removal from the atmosphere-cap
system and is, as shown here, fully capable of terminating any enhanced temperature
regime on early Mars in the absence of any other effects. To put it another way, total
degassed COg inventories of <2 bars and the existence of substantially higher tempera-
tures than present are compatible; total degassed COg inventories of <2 bars, substan-
tially higher temperatures than present, and a regolith qualitatively comparable to the
present one are not compatible.
More recent calculations (6) have shown that when reduced solar radiation is
taken into account ~5 bars of COs is necessary to raise average surface temperatures
oi. Mars above freezing. Calculations are presently under way to determine the ef!ect
of regolith adsorption in the evolution of this large amount of atmospheric COg.
REraRENCES
(1) J.B. PoUack and D.C. Black (1982). [carus, 51. 169-198. (2) M.I. Hoflfert, et al. (1980).
Icarus, 47. 112-129. (3) A.P. Zent, et al. (1986), MECA Symposium on Mars: Evolution
of its Ovmate and Atmosphere . 109-111. (4) E.V. Ballou, et al. (1978). Nature, 271. 644-
845. (5) R Kahn (1985). Icarus, 82. 175-190. (6) J. Kasting (1988). MECA Symposium
on Mars: Evolution of its dimate and Mm.Qsphers, 50-52.
162
CHAPTER 6
SPECTROSCOPY AND REMOTE SENSING
OIIBIZED BASALTS 01 TIE SUgFAIS OF lElUS: COMFOSITICIIAL IMPLICATIOHS
OF mASimi SFECTIAL FIOFEITIIS - C. M. Pieters, ¥. Patterson, S. Pratt, J. W. Head.
(Deft, of Ge©l. Sci., Brm& Uaiv., Pfovideace, 11 D2912) sad J. Garvla (MASA, Goddard)
Mtr®ime^e&: B^tk the reHeclaace pmperUm ajid the gmskems^ of a f@w areas 011 the
¥eiiiislaffl surface have bsea mtasured by tfee »ries ®f ¥esera landers. Tliesi data, coyf ted
witk MorMory refleclaace ia®asur@ffl@ffits of rocks n&d mmernk at the liigli Umpet@Jtarm
chijracteristic of the Yemuslafl surface, are used to mfer tie ©lidatioii sMe ©f iron. in. surface
ffluerais based oa ob^ired s|»e€tral cliaracterlslics.
leflectaace properties of the surface of Venus were meisured by the Venera 9 and !0
landing crafts which contaified two wide angle photometers, one oriented away from the
surface and the other oriented toward it, each with five spectral channels covering the
spectral range from about 0.3 to 1.0 pm CD . The Venera 13 and 14 landers contained
multispectrtl scanners tikU obtained color panoraffia images of the surface at three
wavelengths in the yisible C2). Other eiperiments on ¥enera 13 and 14 measured the
geochemistry of surface material demonstrating that the surface is basaltic in general
character with a FeO content of approximately 9% (3). The mineralogy of the surface is
unknown. The oxidation stale of surface material Is Mm not certain, although a reducing
environment has been suggested from indirect evidence (4). A better description of the
current surface environment and mineralogy will set limits to the possible geochemical
Interactions h®W®®ik surface and atmosphere, essential information for understanding the
recent evolution of the planet.
Sfeetud l®fl@€ls»€@ Ff ®f @rli@s »f tk® Siirfs€@ ®f ¥sffiiis: Five band spectra of the
surface obtained by Veneras 9 and 10 are shown in Figure I (from 1). The surface at these
locations is dark and without significant color In the visible, but exhibits a substantial
increase la reflectance beyond 0.7 urn. Possible purees of error in the near-Infrared have
been seriously considered, and are not expected to alter these measurements substantially.
Information from the Venera 13 and 14 panorfmas indicate the surface is similarly very
dark in the visible (2). Further processing of the panoramas for spectrd information <6)
has shown both the rocks gM soil to be essentially grey (to i 10%) in these visible channels
(wavelengths shown on Figures I and 4). Althougli examination of the spectral character of
unknown phases expected for this €0^ environment have not been completed (e.g. S),ferri€
oxides in oxidised basalt would be distinguished by a characteristic ferric absorption edge (7)
normally observed near 0.55 pm, as In Figure 2. The combined data for the surface of Venus,
however, are not consistent willi the typical room-temperature spectra shown in Figure 2 of
either unoiidiEed bamlls Iwhlch exhibit dari^ reMively fMi continua beyond 0.7 pm,
commonly with superimposed ferrous absorption bands! or oiidked basalts (which derive
their characteristic 'red' coloration from the ferric absorption edge ji&sr 0.55 MbA.
ILmisralaiy l@fl@€tm£€@ M@asiBr@A@iils at BIgli Ts»$@rm£isr@s; The strength, and
often energy, of absorption features is known to be affected by temperature. Many
measurements have examined low temperatures (7,8) with application to Mars and asteroids.
Significant variations also noted at elevated temperatures (e.g. 9) Indicate that only
high-temperature spectra should be used for interpretation of Ihe Venera data, decent
laboratery (lELAfi) mea^remenls have shown this to be critically true for the ferric oxides:
the characteristic idbsorptlon edge moves more than lOOnm to longer wavelengths from 25 to
500'C (Figure 3). This is most likely due te broadening of intense charge transfer
absorptions at shorter wavelengUis.
MsciBssi®s asi CemclBsiens: Venera reflectance data are compared In Figure 4 with
high temperature spectra of the same basaltic materials in Figure 2. The dark. Hat
ufioxidlied basalts are still inconsistent with the Venera data in the near-infrared. Baltic
material with a ferric component, however, would satisfy both the Increase in reflectance
beyond 0.7pffi as well as the dark, relatively colorless character In the m%M®. ¥@ therefore
conclude thai basaltic surfaces of Venus represented by these measurements either contain
minerals with uncommon characteristics, or, more likely, are relatively oiidiEed,
165
Fieters, C,
Use of the KLAB faciiity at Btmn University (supported by NA6W-7«) is ratef«»y
ackfKswIedged. Disoissicms wllh Sm4et coHeagws from Uie Vern^sky institute. IKI, and 60SNIZIPR as part of a
more detailed joint sfialysis of Vei^ra Imder color Images WW's most Ivelpful ^Ing Uiis sUidy.
Mefertncas. I) Economov et al. 1980, /csrvs, 41. o65: 2) Selivanov et ai. 1983, KOStt. I5SLED 2h
no.2. p 176, pl83; 3) Surfcovetal .. 1984, PLPSC 14. JGR.89. B343,- 4) Florensicvetai . 1983, LPS >CIVp203;
1983 KOStl. ISSLED. 2f. no.3, p351; 5) Pieters et @l.. 1985, Bull. Am Astr. Assoc. 17. no 3,p722; 5) ^
Gooding. Plsnet^y Geology reports (ynpubiished); 7) Horris et al. 1985, J6R.90, no. B4. p3l26; 8) Siiwr
andRoush. 1985, y^i?, iniress; 9) Osttorne et ai .. 1978. PLPSC 9th, p2949; Shanklandeta! .. 1979. J6R. ff4.
no.B4. pl603; Parkin etai.. 1980. LPS X!. p854.
Bandpass (half haight) i^ifHTs 1 . {left] i^flect^Ke properties of
U^ Venuslan surface at the Venera 9 and 10
landing sites (from Ecmiomov et al., 1980).
The bandpasses for the Venera 13 and 14
caimras are shown at Uie botlim of Use
figure.
O.S 0.6 0.7 0.8 0.8 1.0 1.1 1.2
Wavelength Cpml
Fiffflr® 2 Irighil. Reflectance ^ectra (0.4 to
1.6 m) of typical basaltic materials at 25X.
The Basalt' is a Iholeite from Taos, NM; Ifie
three cinders are from nauna Kea, HI. The l^ed
Cinder'. Tiaroon Cinder', and Hemalil® exhibit
Ite ch^acterlstlc abs<rption e<^ near 0.55 |im
due lo ferric iron.
WAVELENGTH IN MICRONS
Fifyrs 3 lleftl. ReHeciance spectra (0.4 b
0.8 Mm) of the ferric oxide hematite at
temperatures ranging from rsmn temperature
(^X) to 500X. The trends observed are
reversible. iNote wavelength scale is
expwided compared to Figs 1,2,4.1
WAVELENGTH IN MICRONS
VARIATIONS OF MARTIAN SURFACE ALBEDO: EVIDENCE FOR
YEARLY DUST DEPOSITION AND REMOVAL
Philip R. Christensen, Dept. of Geology, Arizona State University, Tempe, AZ, 85287
Dust deposition and removal is an important process in the evolution of the martian surface.
Mars has been observed to have variable surface markings from the earliest telescopic views of the
planet. These changes have since been seen to be related to aeolian activity, primarily through the
reworking of bright dust deposited following major global dust storms (1,2). Viking Infrared
Thermal Mapper (IRTM) observations of albedo have also revealed significant changes in surface
brightness through time, again primarily associated with major global dust storms (3,4). All of
these observations indicate that there is a significant amount of dust that is deposited during the
decay of global storms which is subsequenSy reworked and redistributed. The purpose of this
study is to determine the degree, spatial distribution, and timing of the deposition and removal of
dust-storm fallout, and to relate tiie current patterns of dust deposition and removal to the long-term
evolution of the martian surface.
A model has been proposed (5) for the development of regional dust deposits that form
through the preferential accumulation of dust-storm fallout into specific northern hemisphere
regions. In this model, dust is deposited uniformly during the decay phase of each major storm,
but is subsequently removed only from regions that are seen today as classic dark areas. Thus,
dark regions remain unman tied by dust, whereas bright regions have developed a 1-2 m thick
mantle of fine, bright dust (5). This model can account for the high thermal inertia (coarse)
material observed in dark regions, together with their relatively high rock abundance (6), and low
albedo. Conversely, bright regions have fine particles (5-40 ^im) and fewer exposed rocks,
presumably due to mantling of the coarse material by dust.
In order to directly observe the seasonal changes in surface brightness associated with dust
deposition and removal, the albedo of specific regions in both hemispheres has been determined
through time. The IRTM data were collected into 1° latitude by 4° longitude bins, at 3 hour
intervals for each 10° of L^. Using these data, the albedo changes for a given area have been
investigated from the beginning of the Viking mission (L^ 84°), through the first (L^ 190-240°)
and second (L^ 270-340°) global dust storms that occurred in 1977. Global data are available
through Lg 120° of the second year, allowing a year to year comparison of surface albedo.
The albedo variations as a function of season are shown in Figure 1 for representative
bright and dark regions. All of the areas studied show a marked increase in brightness associated
with the two global storms, due primarily to the presence of dust in the atmosphere. The increase
in brightness, even for bright regions, indicates that the albedo and scattering phase function of
suspended dust varies from dust on the surface. The maximum brightness at the peak of the
second storm was nearly equal for most bright and dark regions, indicating that the atmospheric
dust was optically thick. For some dark regions, however, such as Solis Planum, the albedo
remained relatively low even at the height of the storm activity, suggesting that the atmospheric
dust was not globally uniform nor well mixed. Many areas show a non-uniform decrease in
brightness during the decay phase, again suggesting spatial variations in dust load and non-
uniform mixing, possibly due to episodic injection of dust into the atmosphere locally (7).
The albedo of most regions had returned to the pre-storm value by L^ 355°, indicating that
the atmosphere had cleared to pre-storm levels by that time. This conclusion is supported by
Viking lander observations, which show that the opacity over the two lander sites had decreased to
167
pre-storm levels by Lg 360° (7). Therefore, surfaces that remained brighter after Ls 360° than they
were prior to the two storms are thought to be covered by a thin layer of bright dust fallout.
The distribution of surfaces that remained bright following the storms, and those where the
surface quickly returned to its pre-storm albedo follow a consistent pattern. The albedo of bright
regions, such as Arabia and Tharsis, rapidly returned to pre-storm values, and was close to the
albedo of the previous year (Fig. la). Many dark regions also darkened to nearly their pre-storm
levels by L^ 360° (Fig. lb). This pattern holds particularly well for southern hemisphere dark
regions. This behavior is consistent with the model of deposition described above; in dark regions
the dust is rapidly removed with little net accumulation, whereas in bright regions a dust mantle
already exists so that the deposition of additional bright dust does not affect the surface albedo.
There are several dark regions that differ from the general trends described above and
provide insight into the level of dust activity that occurs throughout the year. Syrtis Major and
Acidalia Planitia are among the few regions that remained significantly brighter at L^ 360° than they
were before the global storms began. These areas did, however, continue to darken with time,
returning to nearly their pre-storm albedo by L^ 120° (Fig. Ic). It is interesting to note that the
albedo of these and some other regions was still shghtly higher at this time than it was the previous
year, suggesting that some dust still remained on the surface. This finding is consistent with
observations at the Viking lander 1 site where dust was deposited following the global storms and
was not removed until over a year later (8). These observations support the hypothesis that Syrtis
Major and Acidalia Planitia act as local dust sources during inter-storm periods, producing
enhanced dust loading in the northern hemisphere (9).
In summary, observations of seasonal changes in surface albedo reveal regional differences
in the deposition and subsequent erosion of dust-storm fallout. Southern hemisphere dark areas
quickly return to close to their pre-storm albedos, suggesting rapid removal of any dust that was
deposited. Northern hemisphere dark regions are brighter post-storm, but gradually darken to pre-
storm levels over a Mars year. In doing so they act as local sources of dust during otherwise clear
periods. Dust does not appear to be removed from bright regions, resulting in the 1-2 m thick
deposits observed today.
References
1) Thomas, P. and J. Veverka, 1979, /. Geophys. Res., 84, 8131-8146.
2) Veverka, J., P. Thomas, and R. Greeley, 1977, /. Geophys. Res., 82, 4167-4187.
3) Pleskot, L.K. and E.D. Miner, 1981, Icarus, 45, 179-201.
4) Lee, S.W., 1986, submitted to Icarus.
5) Christensen, P.R., 1986, /. Geophys. Res, in press.
6) Christensen, P,R., 1986, submitted to Icarus.
7) Pollack, J.B., D.S. Colbum, R. Kahn, J. Hunter, W. Van Camp, C.E. Carlston, and M.R.
Wolfe, 1977, J. Geophys. Res., 82, 4479.
8) Arvidson, R.E., E. Guinness, H.J. Moore, J. Tillman, and S.D. Wall, 1983, Science, 222,
463-468.
9) Christensen, P.R., 1982, J. Geophys. Res., 87, 9985-9998.
168
a
0.^5
0.40
0.35
0.30
0.25
0.20 -
0.15 -
O.IO
I I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I I ' I I I I I I I I I I I I I I I I I I I I I I 1^
RRRBIR
1 1 I I 1 1 I
D
D
A^
D
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 [ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 J 1 1 1 1 1 1
0. 30. 60. 90. 120. 150. 180. 210. 210. 270. 300. 330. 360.
LS
Figure 1. Variation of albedo with season. Open squares represent data from the first year
following the arrival of Viking at Mars, a) Arabia, typical of bright region behaviour.
D
_J
CE
0.15
0.10
0.35
0.30
0.25
0.20
0.15
0.10
I I I I I I J I I I I I I I I I I I I I I I I I I 1 I M I I I ' I I I ' I I I I I I I I I M I I I I ' I I ' I I I I I ' I I ' I I I I I
MRRGRRITIFER SINUS
-^^"""VV^
n
n
I I I I 1 I I I I r I I I I I I I I f I I I I I I I I I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I f I I I I I I I I I
0. 30. 60. 90. 120. 150. 180. 210. 240. 270. 300. 330. 360.
LS
Fig. lb). Margaritifer Sinus, representative of southern hemisphere dark regions.
0.15
0.10
0.35
,3 0.30
a
5 0.25
en
0.20
0.15
0.10
I I I I I I I I I I I I I I I I I M I ' I I I M I I I I M I I I ' I I I I I I I I I I I I I I I I I I I I I ''' I '' I '
SYRTIS MRJ0R
Ha
^
^D
caa
Q,
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1'-i-'i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 f 1 1 1 1 1 1 1 1
0. 30. 60. 90. 120. 150. 180. 210. 210. 270. 300. 330. 380.
LS
Fig.lc). Syrtis Major, showing the long residence time of dust on the surface in northern
hemisphere dark regions following the global dust storms.
169
ON THE SFKCTEAL lEFLECTAMCE PROPERTIES OF MAflEIALS EXPOSED AT THE ¥IKIMG
LMiDIHG SITES
E. Guinness, R. Arvidson, M. Dale-Bannister, McDonnell Center for the Space Sciences, Department of Earth
and Planetary Sciences, Washington University, St. Louis, MO 63130, R. Singer, E. Bruckenthal, Planetary
Geosciences Division, Hawaii Institute of Geophysics, University of Hawaii, Honolulu, HI 96822.
Reflectance data derived from Viking Lander multispectral data were used
to characterize the types of soils and blocks exposed at the landing sites and
to search for evidence of relatively unaltered igneous rocks. Despite past
research efforts [1-5], there has yet to be a comprehensive examination of the
multispectral data that combines testing of the Camera radiometric
calibrations, explicitly removing the effects of atmospheric attenuation and
skylight, and quantitatively comparing the corrected data to reflectance data
from laboratory materials. This abstract summarizes our work Guinness et al.
[6] which reports such comprehensive efforts. Bi-directional reflectances for
blue (0.40-0.52 ym) , green (0.50-0.59 pm), and red (0,60-0.74 pm) channels
were determined for 31 block and soil exposures at the Viking Landing sites.
In order to interpret the mineralogy of the landing site materials, the Lander
data were compared to laboratory spectra of selected materials and Earth-based
telescopic spectra of Mars after the laboratory and telescopic data were
resampled to the Lander bandpasses.
The uncertainty in the Lander reflectance estimates must be well-known in
order to compare Lander data acquired by the four different Cameras and to
compare Lander reflectance estimates to laboratory reflectance data.
Preflight calibrations were tested by analyses of the brightness values
obtained from sunlit and shadowed gray patches mounted on the Landers, which
have well-defined Lambertian scattering characteristics. Except for Camera 2
on Lander 1, the results were consistent with the combined B% uncertainty in
the radiometric calibrations measured in the preflight tests [7], 3%
uncertainty in the analog-to-digital conversion [8], and the uncertainty in
the optical depth value. We concluded that the preflight calibrations for the
color channels are still valid (with the possible exception of Camera 2 on
Lander 1) and that reflectances can be computed with an uncertainty of about
±10% due to radiometric calibration euid analog-to-digital conversion
uncertainties. Partially shadowed RTC images acquired with the three infrared
channels suggested that there have been changes in the calibration constants
of the infrared channels or wavelength dependent changes in the spectral
responsivity functions. Both types of changes were predicted from modeling of
neutron induced degradation of the infrared channels [7]. As a result, the
infrared channels are poor prospects for extraction of absolute reflectance.
More importantly, it is difficult to make comparisons with laboratory samples
since we cannot determine the nature of the wavelength dependent changes to
the responsivity functions. Thus, we limited our analysis of absolute
reflectances to data from the blue, green, and red channels from the 3 Cameras
where the preflight calibration data are still valid. However, it was clear
from the RTC data that voltage versus RTC gray patch reflectance was still
linear for the infrared channels. This suggested that the IR channel detector
sensitivities were still linear, although absolute values and IR spectral
properties were questionable. Thus, the infrared channels were used to
explore relative brightness variations with wavelength, albeit in more
qualitative ways than with the blue, green, and red channels.
Laboratory analogs used for this analysis were chosen primarily for their
compositional and spectral relevance to Mars, based on the interpretation of,
for example, Viking Lander chemical measurements and Earth-based telescopic
reflectance spectra [9-11]. Samples included the ferric oxides hematite,
goethite, and limonite, two Hawaiian palagonitic soils, hematitically altered
170
volcanic cinders, and seven mafic igneous rocks (fresh and weathered surfaces
of each)« Four telescopic spectra were chosen as representative data in the
spectral region of interest, including a bright region and a dark region
observation from both the 1969 [12] and 1973 [13] oppositions. The
comparisons of the Lander reflectance data to laboratory spectral data was
primarily based on spectral shape. The blue, green, and red reflectances can
be used to characterize spectral shape in terms of spectral curvature and
slope. The red/green reflectance ratio divided by the green/blue reflectance
ratio was used to measure spectral curvature. We found that this curvature
parameter was related to ferric oxide mineral type as well as degree of
crystallinity in laboratory samples. Larger values of curvature would, for
example, be consistent with greater degrees of ferric iron crystallinity
because the green channel is sensitive to the depth of the ferric iron
absorption near 0.45-0.54 ym. This measure was also very sensitive to the
spectral difference between mineral phases such as hematite and goethite. In
addition, we used the ratio of the red and blue reflectances to measure the
slope of the spectra. The red/blue ratio will generally increase as the
degree of oxidation of iron bearing materials increases [14-15].
The variations in bi-directional reflectance in the visible and relative
brightness in the infrared for Lander soils and blocks can best be explained
by varying composition (i.e., degree of oxidation, mineralogy, and
crystallinity) among the different samples. The darkest, grayest block
surfaces are consistent with laboratory reflectance spectra of mafic rocks,
thinly coated with ferric iron-rich palagonite. Such a result is consistent
with the low relative reflectance for the dark facets determined from images
acquired with the infrared channels. More rigorous determination of substrate
rock composition (e.g., ultramafic, basalt, or basaltic andesite) requires
complete spectral data in the near-infrared region. Most other block surfaces
are consistent with thin to optically thick covers of palagonitic material
[10], Limited exposures of soil have blue, green, and red reflectances that
are similar to palagonitic material. One soil exposure has a spectral shape
that is similar to a sample of hematitically altered volcanic cinder. The
bulk of the soils sampled and a few block surfaces have spectra with steeper
slopes and larger curvatures than the palagonite analog. In particular, one
reasonable interpretation is that the most commonly sampled soil exposed at
the landing sites may have a greater degree of ferric iron crystallinity than
the palagonite analog. These soils could also be an intimate mixture of
palagonitic material and other materials not seen in end-member form. But, if
intimate mixing occurs, these other materials would need an even greater slope
and curvature than the observed soils, and be presumably even more different
from material seen on block surfaces in terms of the degree of ferric iron
crystallinity.
The distinctly different spectral shape of most observed soils as opposed
to blocks imply that soils at the landing sites have not been produced by
local weathering of blocks. Even if mixing occurs between material weathered
from blocks and a soil end-member not exposed at the landing site, it implies
that soils at the landing sites are not predominantly derived from local
weathering of blocks. A more plausible scenario is one where soils are
created globally by a number of processes that operated at higher rates
earlier in geologic time. For example, Baird and Clark [16] make compelling
arguments that soils are derived from mafic to ultramafic source blocks in
some isochemical process. Palagonitization of basaltic magmas and basaltic
glass have also been proposed as an important mechanism [17-18]. The
materials would then be transported and would accumulate at a variety of
locations. Once deposited, further erosion would sort the material, and
evaporation of fluids would lead to generation of duricrust. At least at the
171
landing sites, blocks would have been steadily added to the site by impact
events and exposed to slow, perhaps steady state weathering generating only a
small amount of soil. This scenario implies that the bulk of the soil at the
sites carries little information about the local geology.
REFERENCES CITED
[I] Buck, F., and others (1977), J. Geophys. Res., 82,1977 ,
[2] 2) Strickland, E« , (1979), Proc. Lunar Planet. ScL Conf. 10th, 1979.
[3] Evans, D., and J. Adams, (1979), Proc. Lunar Planet. Sci. Conf. 1 0th, 1B29-183A.
[4] Guinness, E. , (1981), J. Geophys. Res., 86,7983-7992.
[5] Adams, J., and others, (1986), J. Geophys. Res.. 91 ,8098-8112.
[6] Guinness, E. , and others, (1986), Proc. Lunar Planet. Sci. Conf. 17ih, in press.
[7] Patterson, W. , and others, (1977), J. Geophys. Res., 82, A391-AA00.
[8] Buck, F., and others, (1975), NASA Tech. Memo. X-72692.
[9] Singer, R, , (1980), Lunar and Planet. Sci. XI, 10A5-10A7 .
[10] Singer, R. , (1982), J. Geophys. Res., 87, 10, 159 -10, U^-
[II] Singer, R. , and E. Strickland, (1981), Lunar and Planet. Sci. XII, 999-1001.
[12] McCord, T., and J. Westphal, (1971), AstrophysicalJoumal, 168, Ul-153.
[13] McCord, T., and others, (1977), Icarus, 3 1 , 293-3U .
[14] Evans, D. , and others, (1981), Proc. Lunar Planet. Sci. Conf. 1 2th, 1A73-1A7 9.
[15] McCord, T. , and others, (1982), J. Geophys. Res., 87, 10, 129 -10, U^-
[16] Baird, A., and B. Clark, (1981), /carwi-, 45, 113-123.
[17] Soderblom, L. , and D. Wenner, (1978), Icarus, 34, 622-637 .
[18] 18) Allen, C, and others, (1981), Icarus, 45,3 A7 -369.
172
SPECTRAL MIXTURE MODELING t FURTHER ANALYSIS OF ROCK AND SOIL
TYPES AT THE VIKING LANDER SITES
John B. Adams and Milton 0. Smith
Department of Geological Sciences
University of Washington, Seattle, WA 98195
We have applied a new image-processing technique to
Viking Lander multispectral images (Adams et al ., 1986).
Spectral endmembers were defined that included soil, rock
and shade. Mixtures of these endmembers were found to
account for nearly all of the spectral variance in a Viking
Lander 1 image, thus one spectral type of soil and one type
of rock mixed in various proportions accounted for the
observed "color" differences.
Further work on the Viking Lander 1 image has solved
the problem of the nature of a minor spectral endmember that
we originally identified as being due to secondary
illumination effects. That endmember can now be shown to be
the result of lighting effects (especially near-far/field
spectral phase changes) and the effects of a foreground
misregistration of the three IR bandpass images with the
three visible bandpass images in the original Viking data.
The misregistration of the data is not apparent when
viewing individual bandpass images or color composites;
however it is distinct in the RMS (root mean squared) error
image which displays the accumulated error between the model
image and the real image. The pixels with high RMS error do
not correspond spectrally or spatially to reasonable soil or
rock types, however they force the introduction of another
spectral endmember.
The misregistration was corrected by Dr. E. Guinness at
Washington University (St. Louis). When the spectral mixing
model was calculated for the rectified images using rock,
soil and shade as endmembers the RMS error image showed only
a weak spectral phase shift from near to far-field, and no
additional endmember was required. These results further
emphasize the usefulness of the mixing model for isolating
geologically meaningful spectral variance from that
introduced by the imaging system or by lighting effects.
Analysis of other Viking Lander 1 and 2 images shows
that the spectral phase effect is strong at certain lighting
geometries. In all cases the effect is expressed as a
gradual shift in the reflectance spectra of all objects from
near field to far field. The spatial patterns of these
shifts are unrelated to geological features. However, the
phase effect must be isolated and removed in order to
compare soil or rock spectra in different parts of an image.
173
We have applied the mixing model to other Lander 1 and
2 images using the same spectral endmembers for soil and
rock that fit the data in the Viking Lander 1 image. The
objective has been to test for the presence of other soil
and rock types by analysing the RMS error images. Materials
that appear spectrally different from the endmembers (and
their mixtures) used in the model appear bright in the RMS
error image. Preliminary results show that the same soil and
rock types carry through all images studied at both Viking
sites, however there are important local exceptions where
areas of spectrally different rocks also occur. As discussed
above we can establish that these spectral differences are
not produced by lighting effects or by system defects. An
evaluation of the nature and extent of these new spectral
rock types is underway,
REFERENCE
Adams, J,B., Smith, M.O,, and Johnson, P.E., Spectral
mixture modeling: A new analysis of rock and soil types at
the Viking Lander 1 site, J, Geophys . Res ,, 91 , 8098-8112,
1986.
174
HISINGERITE AND IDDINGSITE ON MARS: Degradation of Iron-Rich Basalts.
Bymi, Roger G., Itepartiaeal of Earth. Atmospheric and Planetary Sciences,
Massachusetts Institute of Technolc^y. Cambridge, MA 02139.
Iron-rich basalts, which erupted onto the surface of Mars from massive
shield volcanoes such as Olympus Mons, ha^e uadergone eitensive alteration
to produce regoiith dominated by Si, Fe, Al, Mg, Ca, S, and perhaps Ns and
HgO or OH". The modal mineralogy of marlian regoiith is believed to be
dominated by clay silicates, such as the smectites iron-rich montmorillonite
or nontronite, coexisting with poorly crystalline and magnetic iron oiides,
and Ca-Mg sulphates or hydroio ferric sulfate minerals (e.g. jarosite) 111.
The phyllosilicate and ferric oilde phases were deduced to be poorly
crystalline |2| in order to account for reflectance spectral profiles of Mars*
surface in the visible-near infrared region Bl. Such phases also constitute
iddingsite, a deuteric alteration product of olivine in basalts, which has been
identified in meteorites believed to have originated from Mars, Paragenetic
evidence summarized here indicates that hisingerite. too, may have formed
during the evolution of martian regoiith.
Numerous terrestrial occurrences of hisingerite, possessing vitreous luster,
conchoidal fracture and pulverizing to orange-brown powder which
resembles the color of martian regoiith, have been documented [4].
Hisingerite is formed by deuteric and late-stage alteration of pyroxenes and
olivine in mafic igneous rocks, particularly those associated with sulfide ore
deposits I5,6|. Apparently, acidic solutions formed during the dissolution of
sulfide mineralization has led to the formation of hisingerite in austs,
fracture fillings and <xx)ling joints of iron-rich basaltic host-rocks. Similar
environments of chemical weathering may also exist on Mars fl,7|.
Hisingerite was onc^ regarded as poorly crystalline iron -rich smectite or
nontronite f4|. However, recent electron microscopy and X-ray studies f5,6|
have revealed hisingerite to have an amorphous or gel structure containing a
disordered array of [FeO^I octahedra and [5104] tetrahedra. It is just this
coordination environment and degree of crystallinity that matches materials
simulating the spectral properties of bright regions of Mars 151. Therefore,
hisingerite and basic ferric sulfate minerals (e.g. jarosite) appear to be major
contributors to remote-sensed reflectance spectral profiles of Mars [8|.
[II Burns, Miwre,320, 55 (1986); (21 Sherman et aJ.. JGR.%1 , 10169 U982I;
131 Singer, Adv. Space ScL^, 59 (1985); (41 BngaUi, Proc. Ini Qay ConJ'.
BoJ(^m 1981, 97-1 10 (1982); (51 Eggleton. Clavs & Qav Minerals. 32. 1
(1984); [6|Shayan, Qays & Qay Minerals. 32.272(1984); [7] Burns. .76}?, in
press; 18] Research supported by NASA grant number NCR 7604.
175
GC^SANS ON MARS: SPECTRAL FEATURES ATTRIBUTED TO JAROSITE.
Burns, Roger-6. (Department of Earth, Atmospheric and Planetary Sciences.
Massachusetts Instityte of Technology, Cambridge, MA 02139.
Summary, jarosites, (K*,NaMl30*)(Fe3*.Al3*)3(S04)2(OH)6, which are present
in terrestrial gossans capping oiidized sulfides associated with mafic igneous
rocks, may also be present in martian regolith. Spectral characteristics of
jarosites, attributed to Fe3* in the visible region and to water or hydroiyl
groups in the infrared region, are also displayed in remote-sensed
reflectance spectra of bright regions of Mars' surface. The occurrence of
jarosite in the regolith would imply that acidic permafrost and sulfide ores
exist beneath the surface of Mars.
Introduction. The Viking XRF eiperiment detected high sulfur and iron
contents in the martian fines 111. These elements are believed to exist as
oxidized SO^^" and Fe3* species, respeaively [21. Although the majority of
ferric iron on Mars is probably derived from chemical weathering of
ferromagnesian silicates, some Fe3* may have originated from oxidation of
iron sulfides present as accessory minerals in iron-rich basalts or as massive
pyrrhotite ores associated with ultramafic igneous rocks. Several hydroso
ferric sulfate minerals could be stabilized at low temperature and pH
conditions in martian permafrost {31. However, as a constituent of martian
regolith, jarosite is the prime candidate because in arid regions on Earth it is
associated with mixtures of poorly crystalline ferric oxides ("limonite") and
silica (jasper) found in gossans. Spectral features summarized here of
assemblages containing jarosites and poorly aystailine Fe(X)H plus silica or
clay silicates match closely profiles measured in remote-sensed reflectance
spectra of Mars.
Spectra. Reflectance spectra of ubiquitous bright dust and soil on Mars in
the near UV-visible region are highly diagnostic of ferric iron {4|.
Distinguishing features include a slope change at 0.5-0.6 |im and a broad
band centered at 0.8-0.9 jtm. A variety of candidate ferric-bearing phases
have been suggested [4,51, and it is generally agreed that Fe3* occurs in
amorphous or poorly crystalline hydrated ferric oxide-silica gel and clay
silicate phases. However, jarosite has similar spectral features in the UV-
visibie region [61. In the near infrared, spectral features between 1.4-1.7
|im, around 2.3-2,4 jxm, and at 2.9 jim have been assigned to H2O and OH' m
phyllosilicates. Again, jarosite speara possess similar features 17]. In
addition, the mid-infrared feature near 9 jim observed in suspended martian
dust 12] also occurs in spectra of jarosites.
176
Formation of krosite. Near-surface oiidalion of sulfides on Earth proceeds
by electrocftemical processes involving oxygenated groundwater and
dissolved Fe3* loos |8,9l Primary pyrrhotite in mafic igneous rcM;ks is
oxidiEed by Fe3* to secondary pyrite well below the water table (FeySj *
6Fe3* = 4FeS2 + 9Fe2*). Nearer the water table, oxidation of pyrite occurs
(4FeS2 + 2H2O * I5O2 = 4Fe3* + 8SO42- + 4H*). leading to dissolved Fe2*, Fe?*,
and SO42- is acidic groundwater, in which complex ferric ions predominate
(e.g. FeS04*, FeOH2*, etc.). Monodispersed sols of hydronium jarosite
(carphosiderite) may form by reactions involving these wmpiex ions
(2FeS04* + FeOH2* + 6H2O = (H30*)Fe3(S04)2(OH)6 ^ 4H*}. or by direct
oxidation of dissolved Fe2* (6Fe2* > 4SO42- + IIH2O + 3/2 O2 =
2(H30")Fe3(S04)2(OH)6 + 4H*). In the presence of alkali metal cations and Al
derived from chemically weathered feldspars in basalt, extremely insoluble
jarosltes are precipitated. Stability relationships [10] indicate that jarosite
compositions as sodic and aluminous as {Ko,5Nao,5)(Feo,5Alo.5)3{S04)2(OH)6 are
stable at pH 5 and 298 ^K. Below the freezing point of water, they might be
stable to pH 6 and form near permafrost on Mars.
Aoolications to Mars . Independent evidence for the occurrence of jarosite on
Mars stems from its suggested presence in shergotlite and nakhlite
meteorites 111,12], The existence of jarosltes on Mars provides important
clues about the inventory of volatiles, degassing of the mantle, evolution of
the atmosphere, and acidity of groundwater and permafrost there I2,3,9|.
The presence of jarosite in the highly colored gossan-like regolith might also
indicate that massive sulphide ore deposits occur beneath the permafrost of
Mars 1131.
References,
111 Clark ei sl,„ M,^7^ 10059 (1982); |2| Clark k Baird, m, 84, 8395
(1979): Clark & van Hart. Icm-m, 45, 370 (1981 ); [31 Burns. Nature, 320, 55
(1986): f4lSiager.JfiV,J.>^£i?.lfe'; 5, 59(1985); 151 Sherman etaJ.,m.%l,
10169 (1982); f6| Hunt & Ashley, Econ. i^l. 74, 1613 (1979): I7| Hunt.
GeopAysjcs, 44. 1974 (1979); f8| Blain k Andrew. Minerals Sd. Eagng, 9.
119 (1977); 19] Burns, ^^i^in press; 11 0] Hladky k Slansky, BuJl Mineral.,
104. 468 (1981):lll] Smith & Steele, Meteoritks. 19, 121 (1984): fl2j
Gooding. Geochim. CosmocJbim. Acta. 50. 2215 (1986); [131 Research
supported by NASA grant number NSG 7604.
177
CHARACTERIZATION OF SURFICIAL UNITS ON MARS USING VIKING OHBITER
MULTI SPECTRAL IMAGE AND THERMAL DATA
M.A. Presley andR.E. Arvidson, McDonnell Center for the Space Sciences, Dept. of Earth cmd Planetary Set.,
Washington University, St. Louis, MO 63130; andP.R. Christensen, Dept. of Geology, Arizona State
University, Tempe, AZ 85287
The topography of the northern hemisphere of Mars is dominated by three
areas of high elevation (up to 5 km above the datum for Elysium and Arabia,
and up to 27 km for Tharsis) separated by areas of low elevation (about 1 to
2 km below the datum). Albedo and thermal properties correlate fairly well
with this regional topography [1]. The high elevation areas have relatively
high albedos (0.32 - 0.33) and low thermal inertias (< 2.5 x 10~^cal*cm~^s~
1/2^-1)^ while the lower elevation areas generally have lower albedos (<
0.15) and higher thermal inertias (5 - 14). In order to examine these
correlations on a more local scale, a detailed study was conducted [2] of the
types and origins of materials exposed in the central equatorial region
(335''W - 15 °W, 10°S - SO^N). The region is part of the heavily cratered
terrain, located on the western edge of Arabia where the topography is
changing from the high elevations of Arabia in the east to lower elevation in
the west. This area was selected because it displays a wide variation in
color, albedo and thermal properties over a small area of the planet and has
good data coverage, relatively free from dust and haze.
Three surficial units can be distinguished in this region on the basis
of spectral reflectance properties determined from radiometrically calibrated
Viking Orbiter color images and thermal properties determined from the Viking
Infrared Thermal Mapper. These units are; (1) a bright red unit that has a
relatively high reflectance in both red (average reflectance of type area =
0.17) and violet wavelengths (0.06) and a low fine component thermal inertia
(modal value of the unit = 2,4) | (2) a dark violet unit that has a relatively
low reflectance in both red (0.10) and violet wavelengths (0.05) and a
relatively high thermal inertia (6.4) | and (3) a brown unit that has a
reflectance that is relatively low in the violet wavelengths (0.04), but
intermediate in value relative to the other units in the red wavelengths
(0.13), and a thermal inertia that is also intermediate in value (4.6).
Reflectance values for the region produce two trends. These trends suggest
that a bright red end-member material is mixing with each of the darker end-
members. The darker units, however, do not mix with each other.
These units can be mapped to contiguous, well defined locations by use
of simple parallelepiped techniques. Material that produces the brighter
reddish trend, at the apex of the two mixing trends, is located in the east
where it corresponds to Arabia and in the very south where it corresponds to
Deucalionus Regio. Dark violet material is located in the south and
southwest where it corresponds to Sinus Meridiani and Sinus Sabaeus. The
dark violet material also exists as dark splotches in large craters within
the Oxia Palus quadrangle, as well as dark streaks associated with the
splotches. Brown material is located in the north to northwest where it is
the dominant unit in the western half of the Oxia Palus quadrangle. In
addition to the contiguous units, bright red material can be found as bright
crescents within the large craters in Oxia, opposite to the dark splotches,
and as bright margins surrounding the dark streaks.
178
Color/albedo boundaries are relatively sharp and distinct in all places j
yet they are not correlated with abrupt changes in regional morphology, as
determined from medium and high resolution Viking Orbiter images, nor do they
exhibit any systematic correlation with topographic boundaries. Crater
statistics, derived from high resolution Viking Orbiter images and Mariner 9
A-frames, indicate that the units are indistinguishable in terms of age, and
that only a relatively minor amount of crater obliteration (< 0.008 ym/year
depositional rate) could have occurred over the past several billion years.
A major objective of the Mars Observer Mission is to map the physical,
elemental and raineralogical characteristics of the surface, for the purpose
of characterizing both the surficial materials and bedrock geology. The
results of this study [2] suggest that the dark violet unit is probably
composed of sand-sized deposits [3]. The bright red unit is composed of
dust, probably a globally homogenized deposit emplaced during dust storms [4
and 1]. The thermal inertia of the brown unit, and the lack of mixing
between the brown and the dark violet units, are consistent with significant
duricrust formation in the brown unit, as suggested by Kieffer et al. [5].
This possibility is supported by Viking Lander color data. Duricrust exposed
at the Viking 1 Lander site occupies the same relative position on a two-
dimensional histogram of the number of pixels as a function of their red and
blue albedos for Frame A168 (VLl, Sol 28), as the brown unit does on a red-
violet histogram from Viking Orbiter Apoapsis color. These data support the
possibility of extensive duricrust formation and exposure in the brown unit.
If there are significant amounts of duricrust within the brown unit,
then the thermal inertia of the unit suggests that the matrix must be
composed of very fine-grained sand or dust. Thus the brown unit may be
composed of fine-grained bright red material that has been cemented to
produce a duricrusted surface. Although the resultant spectra of matrix and
cement would not be a simple linear addition of end-member spectra, since the
cement would probably form an intimate mixture [6] with its matrix, the
spectra would nonetheless be intermediate between those of the cement and the
matrix. A salt that is transparent in the visible, such as Kieserite,
MgSO^'HjO, which has been suggested as a likely candidate at the Viking
Lander sites by Toulmin et al. [7], could not produce a composite spectra for
the brown unit that is lower in albedo than the red unit. Trans-opaque
minerals, such as iron sulfates (Quenstedtite, Fe2(S04)3'10H20, for example
[8]), however, may present some interesting possibilities, and spectral
properties of such compounds should be investigated.
Another possibility is that the brown unit is simply a lag deposit
composed of a mixture of particles with an effective grain size in the range
of fine sand. This is also consistent with Viking Lander observations.
Viking Lander high resolution color images synthesized by Dale-Bannister [9]
show exposures of brown and bright red materials, with the brown materials
underlying the brighter, redder deposits. Only the redder deposits form
tails or drifts, suggestive of eolian deposition. The absence of dark violet
materials in these images is consistent with Viking Orbiter Apoapsis color
data, which show only brown and bright red materials in the region of the
Chryse Landing site.
A global view shows that both the bright red material and the brown
material occupy large regions of the planet, regardless of morphology or
179
inferred geology. Together vith the results summarized here? this
observation implies that these two units are thin eolian deposits completely
decoupled from the underlying bedrock. The higher elevations of the bright
red unit imply that deposition and/or erosion are being controlled by
topography. Data utilized in this study are insufficient to determine
whether the dark violet unit is locally derived from bedrock, or whether it
is also a thin eolian deposit. Earth-based spectra of the dark regions j
however, indicate that Fe^"^ absorptions around 1 um vary with location on the
planet. These are believed to reflect differences in the mafic mineralogy,
primarily pyroxenes and olivines [10]. These variations suggest that the
dark material is derived from local sources. Since the material of the dark
violet unit is the most likely to be derived from bedrock, data collection by
the Mars Observer Mission should concentrate on these areas, and other dark
regions like them. At any rate, a major problem in analyzing Mars Observer
data will be in determining to what extent a surficial material may be
related to underlying crustal geology or to laterally homogenized eolian
deposits.
REFERENCES!
[1] Christensen, P.R. (1986), Regional dust deposits on Mars; Physical
properties, age, and history, J. Geophys. Res., 91, 3533-3545.
[2] Presley, M.A. (1986), The origin and history of surficial deposits in
the central equatorial region of Mars, MA Thesis, Washington
University
[3] Christensen, P.R. (1983), Eolian intracrater deposits on Mars:
Physical properties and global distribution, Icarus, 56, 496-518.
[4] Christensen, P.R. (1982), Martian dust mantling and surface
composition; Interpretation of thermophysical properties, J. Geophys.
Res., 87, 9985-9998.
[5] Kieffer, H.H., Davis, P. A. and Soderblora, L.A. (1981), Mars' global
properties; Maps and applications, Proc. Lutiar Planet. Sci. Conf., 12B, 1395-
1417.
[6] Singer, R.B. (1981), Near-infrared spectral reflectance of mineral
mixtures; Systematic combinations of pyroxenes, olivine, and iron
oxides, J. Geophys. Res., 86, 7967-7982.
[7] Toulmin, P.T., III, Baird, A.K. , Clark, B.C., Keil, K. , Rose, H.J.,
Jr., Christian, R.P., Evans, P.H. and Kelliher, W.C. (1977),
Geochemical and mineralogical interpretation of the Viking inorganic
chemical results, J. Geophys. Res., 82, 4625-4634.
[8] Clark, B.C. (1978), Implications of abundant hygroscopic minerals in
the Martian Regolith, Icarus, 34, 645-665.
[9] Dale-Bannister, M.A. (1986), Synthetic high resolution color slides
from Viking Lander imaging data , Washington University.
[10] McCord, T.B., Huguenin, R.L., and Johnson, G.L. (1977) Photometric
imaging of Mars during the 1973 opposition, Icarus, 31, 293-314,
180
Wz: AI^IMPTIQN ON PALAGOMTE AND THE MARTIAN REGOLFffl
Aaron P. Zent, Fraser P. Fanale and Susan E. Postawko, Planetary GeoscLences Divi-
sion, Hawaii Institute of Geophysics, University of Hawaii, Honolulu, Hi.
If there was ever suflicient COg in the martian atmosphere to engender a
greenhouse effect sufficient to raise the surface temperature above 273 K, most of
that CO2 is missing today.
Pollack et al., (1986) calculated that, even if reduced solar luminosity is
accounted for, 0.75 bars of COg could have provided surface temperatures greater
than 273K at restricted times and places. We consider 0.75 bars to be the lower
limit on the early atmospheric COg abundance necessary for a greenhouse effect.
Known reservoirs of COg do not hold anyTurhere near 0.75 bars. Only 8 mbar is
currently in the atmosphere, less than 10 mb has escaped to space (Pollack and
Yung, 1980), and no more than a few mbar are in the permanent south polar cap
reservoir (Fanale et al., 1982). In the absence of a deep subsuction mechanism, the
remainder must have been lost to space, or be stored in the regolith, either as car-
bonate rock or in the adsorbed state.
Based on laboratory measurements of COg adsorption on basalt and nontronite,
(Fanale et al., 1982), suggested that it is unlikely that the regolith could have
adsorbed more than about 0.28 bars of COg.
The possible scenarios for the evolution of the martian climate are likely to lie
along a continuum whose end members may be defined as follows:
1) The abundant erosional features indeed indicate the presence of fresh liquid
water (eutectic = 273K) on the martian surface. That water was stablized by the
presence of a massive COg atmosphere, at least 0.4 bars of which cannot be
accounted for in known reservoirs, and which presumably exists as carbonate in the
regolith.
2) The erosional activity was accomplished mainly by Martian brines (Brass,
1980) with depressed eutectics. In this case, sapping, and not rainfall is involved,
and the difficulties associated with fresh water erosion are mitigated, but not
removed. A greenhouse may still be required, but less than 0.75 bars would be
necessary, and it may be that the entire COg inventory necfessary for a greenhouse
eflfect of the required magnitude could be accommodated today in the adsorbed
state without recourse to carbonates.
In the interest of determining an upper limit on the adsorptive capacity of the
martian regolith, we recently examined the results of Fanale and Cannon,
(1971,1974) for COg adsorption on nontronite and basalt. There appeared to be a
strong proportionality between the capacity of the adsorbent and its specific sur-
face area. In order to investigate the h3T)othesis that the specific surface area of
the adsorbent could be used to constrain its adsorptive capacity, even in the
absence of mineralogical information, we performed additional adsorption measure-
ments of COg on palagonites from Mauna Kea, Hawaii. We chose palagonites because
they are a weathering product of basaltic glass and form a good spectral and chemi-
cal analog to the Viking soils (Evans and Adams, 1980, Singer, 1982). Weathering
products generally dominate adsorption behavior, and nontronite is now in disfavor
as a Mars analog material because of features in its spectrum which are undetect-
able or absent in Mars' spectrum.
Determining the relative importance of surface area and mineralogy is impor-
tant becasue m situ measurements of specific surface area for Martian materials
181
exist, whereas no determination of mineralogy has ever been made.
Adsorption measurements were made on a Micrometrics AccuSorb 2100E Phy-
sical Adsorption Analyzer. ¥e found that our data were reproducible to approxi-
mately 40% at low adsorpLive coverage, but that precision increased to better than
20% at higher coverage. The desorption branches of the isotherms were reproduci-
ble, provided the preceding adsorption branches were carried up to near satura-
tion. Unfortunately, the high saturation pressure of COg complicated this in higher
temperature data runs. However, the BET surface area, which we calculated for
each isotherm, provides another check on internal consistency. Performing a
least-squares fit on the data, we find that the data can be fitted to a curve of the
form
p^=6¥fT^ (1)
where 5 = 27.93, y = 0.171045. and fi = -1.44675. Here P is in mm Hg, and.T is
degrees Kelvin. The associated error in this equation is i20%, based on the spread
in the data.
Normalizing for the surface area of all three minerals, decreases the uncer-
tainty associated with not knowing the composition of the adsorbent to less than a
factor of three from approximately a factor of 13. We believe that accounting for
specific surface area provides a significant improvement in our estimate of the
capacity of a mafic particulate adsorbent at Mars-like conditions. More impor-
tantly, it provides a description of the response to temperature variation which is
less mineralogy-dependent and more useful in predicting climate change than those
used in the past.
For the purposes of this study, -we therefore assume that an equation can be
written which describes COg adsorption on any basalt or basaltic weathering pro-
duct in terms of temperature, PcOg and the specific surface area of the material,
without regard to mineralogy. We do not apply this equation to temperature and
pressure conditions which differ from the conditions at which the data were gath-
ered. The complete data set for all minerals was used to fit that equation. The
equation which best describes the general COg adsorptive behavior of particulate
mafic rocks and their weathering products is
p« = A«6F'TP (2)
Where Ag is the specific surface area of the material in (m^ g~0. and P and T are as
above. In this case 6 - 5.9629 x 10*. y = 0.364391, and /3 = -3.83415. A check of
this equation indicates that it predicts adsorptive capacity to within a factor of
three for those mafic rocks and weathering products for which COg adsorption data
are available, regardless of precise mineralogy.
If we are granted the indulgence of scooping up a few cubic centimeters of soil
and extrapolating to the entire unconsolidated regolith, we can use Eq.(2) to draw
some constraints on the partitioning of COg in the martian regolith.
Fanale et al. (1982) presented a model of the martian climate in which they
considered regolith adsorption to be the primary reservoir of exchangeable COg.
They presented results based on basalt and nontronite, and described cap forma-
tion and atmospheric pressure as a function of obliquity. We have reconfigured that
model, eliminating explicit assumptions of mineralogy, and substituting Eq.(2) with
Aa = 17 m^ g-^
182
There -ace two independent variables in the current configuration of this model.
One is the amount of exchangeable COg, (i.e. adsorbed, atmospheric, or cap COg).
The other variable is the depth of the regolith, which essentially determines the
total surface ar(;a avaihibk; for adsorption. This model is useful because if we
specify the depth of the regolith that is in diflfusive contact with the atmosphere,
then there is a unique amount of exchangeable COg that is compatible with, 1) the
current atmospheric pressure, and 2) the absence of a significant current polar COg
reservoir. Our model allows us to find that amount of exchangeable COg.
The depth of the regolith in our model is not the depth of crustal fracturing,
but a conceptual construct. It is the equivalent depth of a model regolith, with
specific surface area 17m^g~\ in diffusive contact with the atmosphere. A rela-
tively small thickness of this idealized regolith would be the equivalent, in terms of
total surface area, of a rather extensive real regolith, with sills, blocks, adsorbed
HgO and probably extensive permafrost. Our model indicates that even 500m of this
ideal, highly adsorptive regolith could hold no more than 0.19 bars.
There are reasons to place a lower limit on regolith thickness as well. If we fbc
the total amount of outgassed COg, and make the regolith progressively shallower,
the concentration of carbonate in the regolith must increase, assuming that COg
displaced from adsorbate would remain in the regolith. Eventually, a finite amount
of carbonate should be present at the surface. Nonetheless, there is no spectral
evidence for the existence ot carbonates on Murs. That argues against a massive
carbonate deposit of unconstrained mass as a panacea for mode's involving massive
greenhouse effects early m martian history. Nonetheless, it is clear thai some
reservoir over and above adsorbate is almost certainly necessary.
This model allows us to make some estimates of exchangeable COg abundances
without knowing m^uch more than we do now about martian volatile history. All
atmospheric pressure vs. obliquity caives cross at obliquity 25" and pressure 7.5
mbar, as would be expected since Lhe pairs of independent variables were- c;hosen to
be consistent with current conditions, i^'owever as long as the exchangeable COg
abundance and regolith depth are chosen with those constraints in mind, atmos-
pheric pressure and cap formation as a function of obliquity is only weakly depen-
dent on the specific parameters we chose. The actual mass of the polar cap at low
obliquity of course does depend strongly on total exchangeable COg, but the atmos-
pheric pressure does not at any obliquity.
REFERENCES
1. Brass, G. W., [cams. A2. 20-28, 1980.
2. Evans, D. L, and Adams, J. B. Proc. Lunar and Planet. Sci. ConJ. 11th. 757 - 763.
1980.
3. Fanale. F. P., and Cannon, W. A.. Nature. 830. 502-504, 1971.
4. Fanale, F. P.. and Cannon, W. A. /. Geaphys. Res. 79. 3397-3402, 1974.
5. Fanale, F. P., et al.. Icarus. 50. 381-407, 1982.
6. Pollack. J. B.. and Y. L. Yung, Ann. Rev. Earth Planet. Sci., 8. 425-487, 1980.
7. Pollack, J. B., et al.. Submitted to: Icarus, 1986.
8. Singer, R. B., / Geaphys. Res., 87, 10,159-10,169. 1982.
183
INVESTIGATION OF MARTIAN H2O AND CO2 VIA GAMMA-RAY SPEC-
TROSCOPY
Steven W. Squyres, Cornell University, Ithaca, NY 14853, and Larry G. Evans, Computer
Sciences Corporation, Beltsville, MD 20705.
Earth-based and spacecraft investigations of Mars have shown that H2O and CO2 are the two
most important volatile species on the planet. While major advances have been made concerning
the evolution and present state of H2O and CO2 on Mars, many unanswered questions remain.
The upcoming Mars Observer mission will include in its payload a gamma-ray spectrometer.
This instrument will measure the spectrum of ganmia radiation emitted by the planet. Prom
the measured gammarray spectrum, it is possible to infer a considerable amount of composi-
tional information. Particularly, the experiment will enable study of the spatial and temporal
distribution of H2O and CO2 in the near-surfsice martian regolith.
The presence of H2O can be investigated directly by measurement of the 2.223 MeV gammarray
line of H. Because H is the most important element for moderating the neutrons produced by
cosmic-ray interactions with the surface, it can also be investigated by the direct or indirect
determination of the fast to thermal ratio of the neutron albedo flux. In general, for materials
with high H concentrations, the total neutron flux consists of proportionally more thermal
neutrons and the flux peaks at shallower depths than for materials with low H concentrations.
The neutron fast/thermal ratio may be determined indirectly from the ganuna-ray spectrum.
Prompt capture lines result primarily from interactions of nuclei with thermal neutrons, while
inelastic scattering lines result primarily from interactions with fast neutrons. Some elements,
such as Si and Fe, emit strong lines of both types. One may therefore examine ratios of inelastic
scatter to prompt capture line strengths for these elements, and acquire information regarding
the H distribution that is similar to what one would acquire directly from fast/thermal neutron
ratios. The 2.223 MeV gamma-ray flux is an indicator of the amount of H in the upper few
tens of g cm~^, while the inelastic/capture ratio for Si or Fe is related to the amount of H
in the upper ~ 100 g cm~^ or more. It is therefore possible to obtain information about the
vertical distribution of H. CO2 is more diSicult to detect, but it may be possible to determine
the thickness of a layer of CO2 frost by direct detection of C or by inference from attenuation
of gamma rays from underlying material.
We wish to determine how efiFectively questions regarding the distribution of H2O and CO2 on
Mars may be addressed with orbital gamma-ray data. Our approach is straightforward. We
have identified several unanswered questions regarding martian H2O and CO2. These include:
(1) What is the ice/dust ratio of the polar layered deposits? (2) What is the thickness of the
polar perennial ice? (3) What is the thickness of the CO2 that covers the southern perennial
ice? (4) What is the thickness of the seasonal CO2 frost caps? (5) How much H2O is there
in the seasonal frost cap? (6) How does the distribution of subsurface ice vary with latitude
and geologic material? (7) What is the degree of hydration of minerals making up the naartian
surface? (8) Are there near-surface liquid water "oases" in Solis Lacus or elsewhere? For
each question, we have formulated a simple multi-layer model of the martian surface. In each
model, the martian surface is composed of one or two horizontal layers. K one surface layer
is present, it is taken to be semi-infinite; if two surface layers are present, the upper one has
some variable thickness and the lower one is semi-infinite. A layer may be composed of H2O,
CO2, or soil with a composition like that at the Viking landing sites, or some homogeneous
mixture of two of these materials. In two-layer models in which one layer is a mixture of two
materials, the other layer is always composed of a single material. Each model therefore can
be completely described by no more than two parameters. As an example, the problem of
detection of subsurface ground ice is modeled with a variable-thickness upper layer of pure soil.
184
and a semi-infinite lower layer of intermixed soil and H2O. In this case, the two parameters
are the thickness of the upper layer and the soil/H20 mass ratio of the lower layer. For each
model, we calculate the gammarray spectrum that will be observed by a spacecraft in orbit.
All of the calculations include atmospheric attenuation and emission. We then examine ways
in which the observed spectrum can be interpreted to yield unique determinations of values for
the parameters of interest.
In order to predict the expected gamma-ray fluxes from the surface of Mars and the sensitivity
of the Mars Observer gamma-ray spectrometer experiment, the neutron spatial fsnd energy
distributions in the near-surface materials must be modeled. We use a primary c os^nic-ray
spectrum and resulting secondary neutron source distribution like that for the Moon. We then
calculate the resulting neutron spectrum and distribution for the surfeice and atmosphere under
consideration using the ANISN code. This is a discrete ordinate code that solves the neutron
transport equation by evaluating the flux for discrete directions, positions, and energies. The
code is one-dimensional, and uses 100 neutron energy groups and 100 spatial intervals with a
spacing of 5 g cm~^.
Once the neutron spatial and energy distributions have been determined, we calculate the
discrete line gamma-ray fluxes produced by neutron interactions. For capture lines, the cal-
culated thermal flux is used with the appropriate thermal neutron capture cross-sections and
ganmia-ray yields to calculate the gamma-ray line flux. Inelastic scatter gamma-ray fluxes are
calculated in a similar manner with the inelastic scatter cross-sections integrated over energy at
each depth. Production of ganama rays by natural radionuclides is also calculated. All gamma-
ray line strengths are corrected for attenuation in the martian surface and atmosphere. The
ganuna-ray beickground is calculated from the observed lunar background, scaled to Mars from
gamma-ray transport results.
Combining line strengths, background strength, detector characteristics, and counting time,
we may estimate the quantitative uncertainty of the measurements to be made. For our cal-
culations, we used the area, energy resolution, and efficiency of a laboratory 120-cm~' HPGe
detector, and a counting time of 38 hr. This period corresponds to the total integration time
for a single gamma-ray spectrometer resolution element (about 300 km diam.) near the poles
over one martian year.
We have considered a total of five models:
Model 1: Intermixed Soil and H2O: This model can be used to address problems such as the
ice/dust ratio of the layered deposits, the dust content of the north polar perennial ice, and
the degree of hydration of hydrated silicates. Sample results are given in Figure 1, where we
plot two observed line strength ratios as a function of the mass fraction of H2O present. For
small amounts of H2O, optimum results are given by the ratio of Si inelastic capture to prompt
capture line strengths, while for larger amounts of H2O, optimum results are obtained by using
the ratio of H (2.223 MeV) to Fe (capture). For a simple model such as this, it should be
possible to determine the H2O content to a high degree of accuracy.
Model 2: Soil over Soil + H2O: This model consists of two layers: a pure soil layer of variable
thickness over a semi-infinite layer of soil + H2O with a variable H2O mass fraction. This
model can be used to address the distribution, depth, and concentration of ground ice (or
liquid). Figure 2 plots Si(inelastic/capture) vs. H/Fe(capture). Solid curves are contours of
constant H2O mass fraction in the lower layer, and dashed curves are contours of constant
thickness of the upper layer. For data falling in the region where the curves are separated, it
will be possible to determine both the thickness of the upper layer and the ice mass fraction of
the lower layer.
185
Model 3: H^O over Soil + H2O: This model can be used to address the thickness of the north
polar perennial ice, which lies atop the layered deposits. Calculations like those in Figure 2 show
that unique solutions are possible, but the uncertainties are large. However, if we assume that
we know the ice fraction of the lower layer a priori, (as might be possible from an independent
measurement of the ice content of exposed layered deposits) determination of the ice layer
thickness cjin be improved very substantisdly.
Model 4: CO2 over Soil + H2O: This model can be used to address two problems. In the
extreme of very high H2O mixing ratio in the lower layer, it can be used to investigate the
thickness of tke CO2 layer atop the south polar perennial ice. In the extreme of a very low or
zero H2O mixing ratio in the lower layer, it can be used to address the problem of the thickness
of the seasonal CO2 frost cap. Again, unique solutions are possible, but the uncertainties are
large in some cases. However, if we know the H2O content of the lower layer a priori (as would
be the case if it were determined when the CO2 layer was absent), uncertainties again can be
very substantially reduced.
Model 5: CO2 + H2O over Soil: This model can be used to address the H2O content of the
Seconal frost cap. Unique solutions Me found for both adjustable parameters, with small
uncertainties.
These results indicate that the Mars Observer gamma-ray spectrometer will be a very powerful
tool for investigating the distribution and stratigraphy of volatiles on Mars. It is important
to note that the results here do not include the substantial additional information about the
neutron energy spectrum that will result from the GRS's neutron mode. With this mode, it
should be possible to determine fast/thermal ratios directly, providing an independent check
on the gamma-ray results, and increasing the overall certainty in the results.
10-
— I 1 1 r
INTERMIXED S01L + H20
38 Hr.
10
en
(/)
z
tij
I-
z
>
I-
<
UJ
a:
1.0
Of
50g/cm'
ZOg/cm'
SOIL OVER (SOIL+HjO)
38 Hr
Sg/cm*
-Percent HgO
-Soil Thickness
I
. ...1
icr'
iO"
10
tO^
10^
RELATIVE INTENSITY H/Fe (capture)
Figure 1 (left)
Figure 2 (above)
too
PERCENT H,0
186
HEFLECTANCE SPECTRA OF MAFIC SHICATES AND PHYIIDSMCATES ffiOH .6 TO 4.Q[im:
Ted L. Roush, Robert B. Singer, and Thomas B. McCord, Univ. Hawaii, Planetary Geosciences Div.,
Hawaii Inst. Geophysics, 2525 Correa Rd., Honolulu, HI 96822
IN TM)DUCTION Spectral reflectance, obtained by spacecraft and earth-based telescopic
observations, is the dominant method of collecting mineralogical information concerning the
surface properties of solar system objects. Our ability to interpret this data, in terms of the
identity and abundance of minerals present on a surface, is important for addressing more gen-
eral questions concerning the origin and evolution of that body as well as the solar system as a
whole. The reflectance of geologically important materials in the 0.3 to 2.5/Lan wavelength region
has been the subject of intense laboratory research for many years(l,2,3,4,5). However, there
has been a paucity of research concerning the reflectance behavior of geologic materials in the
2.5 to 4.6/im wavelength region.
The mapping spectrometers on future spacecraft missions, such as Galileo and the Mars Orbiter,
are designed to return spectral reflectance data from 0.7 to 5.2fj.xn. Thus, it is important for
interpretation of such data to expand laboratory reflectance measurements to include the
longer wavelength region. The study presented here provides empirical laboratory data concern-
ing the reflectance behavior of geologically important materials in the 2.5 to 4.6jLtm wavelength
region.
EXPERIHENTAL HETHOD For this study two infrared, cooled, circular variable filter (CVF) spec-
trometers with InSb detectors were used and the combined wavelength coverage spanned the .65
to 4.5fim wavelength region. In both instruments the detector, filter, and other internal com-
ponents are operated at liquid nitrogen temperature (77 K). The spectral resolution (AX/X) of
both CVF's is about 1.5%.
The mafic samples were chosen to represent minerals which result in mafic to ultramafic assem-
blages. All the mafic samples were dry sieved to the OSfj-ra grain size. The phyllosilicates were
chosen to represent a variety of structural types and amounts of natural hydration states. All
phyllosilicate samples were dry sieved to the 38-45/zm grain size. Additionally, a palagonite was
included in the sample suite since this material has been sii^gested as a Mars soil analog
material(6,7). The 2-3jj.m grain size of the palagonite was separated by liquid suspension in
methanol.
All samples were placed in a furnace prior to being situated in an environment chamber. The
environment chamber was flushed with inert nitrogen gas during the entire data collection and
heating sequence. The samples were allowed to equilibrate in the nitrogen environment for
thirty minutes before room temperature spectral data were collected. The samples were then
heated at 200°C for thirty minutes and subsequently allowed to cool to room temperature before
additional data were collected.
BACKGROUND Electronic tran^tions of the first transition element series, especially the petro-
logically significant cations Fe and Ti , restilt in absorptions in the visible and near-infrared
(0.3 to 2.5/Ltm) region of the electromagnetic spectrxim(8), and are due to ions located within
specific crystallographic sites. Hence, from the spectra direct information is obtained concern-
ing the chemistry and crystal structure of the material.
For the silicates and of this study, the most important species which have vibrationally induced
absorptions are the HoO molecule and the OH' ion. The HpO molecuJe has strong absorption
features located near Z.9fj.m (0-H asymmetric stretch), 3.0/xm (H-Q-H bend overtone), and 3.1/im
(O-H symmetric stretch)(9). Overtones and/or combinations of these bands can occur near 0.94,
1.135, 1.38, 1.45, and 1.88^tm(9). There is a fundamental of the OH' group located near 2.75/xm
and the first overtone is located near 1.4/.tm(9). Additionally, bands in the 2.2 to ZA/Mca region
involve a combination of the OH' fundamental with a X-OH bending mode (where X is usually Al or
Mg)(9).
DISCUSSION The results of the spectral measurements for the maflc silicates are presented in
Figures 1 to 3. All spectra, except the plagioclase, exhibit electronic absorptions located near 1
and /or 2fj.ra which are due to Fe located in octahedral crystallographic sites. These absorp-
tions remain unchanged after heating. Additionally, all spectra have an absorption centered
near 3;iim which is due to HgO. This absorption changes in intensity, and in some cases shape,
after heating. ¥e believe these changes represent the removal of physically adsorbed HpO from
the sample. The continued presence of the 3/.im band band in the spectra of all samples after
heating implies that HpO bearing contaminants are present, or alternatively, the samples
187
contain fluid inclusions.
The results of the spectral measurements for the phyllosilicates are presented in Figures 4 and
5. Kaolinite and pyrophyllite have types of structures which typically accommodate little
cationic substitution and hence do not incorporate interlayer HgO(lO). Before and after heating,
the spectra of both samples exhibit strong absorption bands located near 2.75^4m, due to OH",
and the spectra after heating show minor changes due to the removal of physically adsorbed
HgO. This spectral behavior is illi:istrated in Figure 4 for kaolinite.
The Na-Montmorillonite (Fig. 5) has a structure which has abundant cationic substitution and its
open nature results in the presence of interlayer HoO(lO). The spectrum of the sample before
heating exhibits a broad 3fj,in band and narrower but pronounced 1.9jLtm band both due to HpO
and bands located near 1.4 and 2.8fMn both due to OH'. After heating, the spectrum exhibits the
effects of removing the physically adsorbed, and possibly some of the interlayer, HgO. The 1.9
and S.OjLim bands have decreased in intensity, while the band near 2.8/um has become more pro-
nounced.
The palagonite spectrum (Fig. 6) before heating shows a strong, broad S/xm absorption and a
weak absorption near l.Q/Lun, both indicate the presence of HoO. After heating, the spectrum
shows virtually no l.d/Mn band and the 3/j.m. band although slightly reduced in intensity still
remains very prominent. We believe the physically adsorbed BIoO has been removed, so the con-
tinued presence of the 3/xm band indicates abundant HpO. Tne spectral behavior of the S/xm
band is reminiscent of the plagioclase.
CONCLUSIONS This study has provided valuable spectral reflectance information about mafic sili-
cates and phyllosilicates in the 2.5 to 4.6jum wavelength region. In this wavelength region we
have shown that the reflectance of these materials is strongly affected by the presence of HpO
and OH . Therefore, the identification of these absorbing species is greatly enhanced.
The refiectanee behavior of mafic silicates beyond 2.5jj.m can provide additional information
about the chemical composition of the pyroxenes by revealing the long wavelength edge of the
clinopyroxene absorption band located near 2/.tm. For Mercury, the Moon, Earth, Mars, and the
asteroids this provides additional informatipn which enhances our ability to map changes in the
mineral composition and abundance across the surface of a planetary body.
The reflectance of phyllosilicates beyond Z.5fj,m provides additional information concerning the
potential for identifying compositional and water content variations of these minerals. Since OH
is structurally boxind to octahedrally coordinated cations in phyllosilicates, the substitution of
various cations should result in the shifting of absorptions due to the OH'-metal vibrations. Addi-
tionally, absorptions due to interlayer HpO can be expected to vary as a function of amount of
water available on the siirface. In the case of Mars, a map of water content variation could iden-
tify possible sources and /or sinks of weathering products on a global scale and allow the study of
water transport on the surface.
The spectral reflectance behavior of the palagonite beyond 2.5/i,m implies that the water con-
tained within the sample is not composed solely of physically adsorbed water. Thus, the palagon-
ite, if located in the desiccated Martian environment, could provide an exchange mechanism for
atmospheric water vapor by trapping and releasing physically adsorbed water. The palagonite,
during its initial formation, also provides a more permanent mechanism for removal of atmos-
pheric water from the Martian environment.
ACKNOWLEBGEHENTS This research was supported by NASA grants NSG 7590 and NSG 7312.
KKfKHENCES l)Hunt, G.R. and J.W. Salisbury, Mod. Geol., 1, 283-300, 1970. 2)Hunt G.R. and J.¥.
Salisbury, Mod. Geol., 2, 23-30, 1971. 3)Adams, J.B., /. Geo-phys. Res., 79, 4829-4836, 1974.
4)Singer, R.B., /. Qeophys. Res., 86, 7967-7982, 1981. 5)Clark, R.N., /. Geophys. Res., 86, 3074-
3086, 1981. 6)Evans, D.L. and J.B. Adams, Proc. Lunar Hanet. Sci. Conf. 11th, 757-763, 1980.
7)Singer, R.B. /. Geophys. Res., 87, 10159-10168, 1982. 8)Burns, R.G., MneralogiccU J^plications
of Crystal Field Theory, 1970. 9)Hunt, G.R., Geophysics, 42, 501-513, 1977. 10)Deer, W.A., R.A.
Howie, and J. Zussman, Jin IntrodvcOon to the Rock Forming Minerals, 1966.
188
' ' 1 '
' ' ' 1 '
1 ' '
' ' 1 '
-
~-
jf^^Ste**
.^
^J
^
—
-
\
n
—\
^
■^^^
^ '
-
V
^b
^5e I
-
I t 1 .
1 1 r
1 > 1 1
1.0 S.O 3.0 4.0
Wavelength Cum)
Figure 1. Reflectance spectra of olivine Cfogg) before (b) and
after (a) heating. The absorption centered near of lyxm is due to
the presence of Fe and is unaffected by heating. The decrease
of the absorption located near S/im upon heating is due to the
loss of physically adsorbed HgO.
Wavelength (Aim)
Figure 4. Reflectance spectra of kaoUnlte (A]oSigOg(OH) J before
(b) and after (a) heating. The spectra exhibit OH' vibrational
absorptions near 1.4. 3.2, and 3.75p.m. The spectral changes due
to heating indicate the loss of physically adsorbed HgO.
o.a
- --
'- ^ r\ A -
0.6
0.4
.a
hfv^
a ^.
b j
'. . . 1 , , , , 1 , , , , 1 , , , ,"
-
—I T i 1
-
.^^^
—
-
a
, 1 .
bN
i
a
2.0 3.0
Wavelength (jum)
1.0 2.0 3.0
Wavelength (.urn)
Figure 3. Reflectance spectra of orthopyroxene CEngg) before
(b) and after (a) heating. The absorotions located near 1 and
S>j,m are due to the presence of Fe^ and are not aflected by
heating. Because this sample contains tremollte as a contam-
inant, after heating the S/an band depth decreases, due to the
loss of physically adsorbed HgO, and the relatively narrow
absorption located near 3.75>im, due to OH', is more pronounced.
a.o 3.0
Wavelength Cum)
Figure 3. Reflectance spectra of plagioclase (Angg) before (b)
and after (a) heating. The spectra of this mineral exhibit the
strongest S/im of all the maflc silicates due to the presence of
fluid inclusions in the sample. Since the HgO in the fluid inclu-
sions is not easily lost upon heating, the minor spectral changes
indicate solely the loss of physically adsorbed HgO.
Figure S. Reflectance spectrum of Na-Hontmorillonite (nomi-
nally, ()iCa,Na)Q 7(Al,Mg,Fe)4[(Si,Al)g03(,](0H)4nHg0) before (b)
and after (a) heating. Before heating uie spectrum of this sam-
ple exhibits absorptions due to HgO near 1.9 and 3.0>im and OH'
near 1.4 and 3.3>an. After heating the spectrum exhibits the
effects of removing physically adsorbed, and potentially inter-
layer HgO, by the reduced intensity of the 1.9 and 3.0;an bands
and the 0H~ band near S.B/im is more pronounced.
-'
' 1 '
' ' ' 1
' ' ' ' 1 '
' ' ' 1
• '.
-
-_
_
-
b
a
-
-
*«i^^i—
^"^
-
: %
a
^1
a
»-i*»-s -
-
I .
^f
~^»=is _
-
, 1 ,
, . , 1
. , ,Wi^
<^ , 1
1.0 2.0 3.0 4.0
Wavelength Cum)
Plgure 6. Reflectance spectrum of palagonite before (b) and
after (a) heating. Before heating the spectrum of this sample
exhibits a weak 1.9fjm band and a strong 3>im band both indica-
tive of HgO. After heating the spectrum exhibits a very weak
1.9;im band implying removal of HgO, yet the 3;im band remains
prominent. This spectral behavior is very reminiscent of the pla-
gioclase.
189
SPECTRAL EPTECTS OF DEHYDRATION ON PHYLLOSIIJCATES; E.A. Bruckenthal
and R.B. Singer, Planetary Geosciences Division, Hawaii Institute of Geophysics, 2525
Correa Road, Honolulu, HI 96822
INTROUCnON Six phyllosilicates have been progressively dehydrated under con-
trolled conditions in an effort to study the spectral effects of their dehydration.
Justification for the study may be found in both terrestrial and planetary soil science
applications. For example, we know through spectroscopic observations of Mars and
several asteroidsl.S that their surfaces contain ubiquitous hydrated silicates. However,
due to the anhydrous nature of the extraterrestrial environment, one would not expect
these soils to exist in the same hydration states as they do on Earth. Furthermore,
spectroscopy provides a tool which probes the internal structures of these clays as
they undergo changes. Thus, the spectra obtained at each level of hydration provide
information which may be used in future spectroscopic observations of other planets,
as well as a data set which complements the existing body of terrestrial soil knowledge.
SAMPLES AND EXPERIMENTAL METOODS Samples were chosen to represent a range of
both crystalline and X-ray amorphous soil components. The four crystalline clays in-
clude dioctahedral and trioctahedral endmembers of t-o and t-o-t type clays, as well as
species characterized by extensive substitutions in their tetrahedral and octahedral
layers. Serpentine, Mg2SiQ0c(0H)^, represents an endmember (negligible cationic sub-
stitution) tricotahedtal t-o clay, while talc, Mg3Si^0<Q(0H)2, is an endmember trioc-
tahedral t-o-t clay. Ca-montmorillonite was chosen as a dioctahedral smectite, while
saponite represents the trioctahedral sm.ectite group. Both smectites have the general
formula, i^ Ca,Na)p ,7(Al,Mg,Fe)^(Si,Al)gppo(OH) . nHgO.
The final two clays studied are Big Isfand palagonites ~ X-ray amorphous, low tem-
perature alteration products of basaltic volcanic glass. The two differ in that the Pahala
Ash is more completely weathered, more incipiently crystalline, and significantly more
hygroscopic than the Mauna Kea palagonite.
Each clay was ground and dry-sieved to <38 yu.m, and then equilibrated in a high
humidity {>95%) environm.ent. An environm.ent chamber which allows spectropho-
tometric observations at non-ambient conditions was used in conjunction with a
custom-built furnace. Each sample was heated to eight elevated temperatures (120,
160, 200, 250, 300, 400, 600, and 750° C.) in an inert, dry nitrogen atmosphere. Spectra
were obtained at room tem.perature and after heating to each of the aforementioned
target temperatures.
BACKGROUND In order to understand the nature of clay dehydration, a brief introduc-
tion to the different species of hydration is helpful. The four species of hydration which
exist include three forms of adsorbed water and one form of hydroxyl ions.
The first of three types of adsorbed water is found in pores, on surfaces, and
around the edges of particulate samples of all clay minerals, regardless of type. It con-
sists of loosely bound HgO molecules held only by Van der Waals' forces and is called
physically adsorbed water.
The second and third forms of adsorbed water, found almost exclusively in t-6-t
layer clays, refer to those HoO molecules which are bound more tightly to a clay
mineral than physically adsorbed water. Although capable of being removed at low
temperatures, these molecules are bound to the basal surfaces of a clay by forces of a
more electrostatic nature. This species will be referred to here as surface-bound wa-
ter. Often surface-bound water grades into physically adsorbed water at some distance
from the basal surface.
Very similar and difficult to distinguish from surface-bound water are those HoO
molecules bound to cations which have been adsorbed between layers in smectfte
clays. A distinction must be made between this cation-bound water and surface-bound
water because researchers have shown that cation-bound water is held much more
tightly, and driven off at much higher temperatures, than surface-bound water-^'''.
190
The final form of hydration consists of those OH molecules which comprise por-
tions of the clay structure through their presence in an octahedral layer. This species
of hydration will be referred as structural hydroxl ions, and its loss from a clay as
dehydroxlation.
INTERPRETATION The plots which follovf illustrate spectral changes observed with in-
creasing temperature for three of the six samples investigated. Only those tempera-
tures at which significant changes occurred are shown. Numbers in bold type represent
the temperatures associated with each sample.
Talc
WavQlcngtri Cum]
Spectral features to be noted in the room temperature (22° C.) spectrum above in-
clude the following: 1) a sharp, 1.4 /zm band due to both OH and HgO 2) a broader 1.9
(xm HgO band 3) a broad, shallow 2.2 fj,m cation-OH feature 4) very sharp, well-defined
2.3 and 2.4 yiim Mg-OH bands 5) a strong OH fundamental near 1.8 /Ltm, and 6) a broad,
deep 3.1 /zm HgO absorption. Upon heating to 120°C., physically adsorbed water contri-
buting to the 1.9 and 3.1 /im bands is driven off, dramatically lessening the strength of
both features. By 400°, more tightly held water molecules have begun to leave the
structure, reducing 1.4 fxm band depth, completely eliminating the 1.9 fim band, and
further reducing the 3.1 /u.m band. Limited cationic substitution has attracted surface
and/or cation bound water molecules to a few interlayer sights, and suflicient kinetic
energy is now available to begin removing them. Most likely some of this water is rein-
corporated back into the structure to cause the oxidation which is apparent as a slope
change between 0.65 and 1.8 /im. Upon heating through 750° C, talc continues losing
its most tightly bound HgO, as shown by the continued decrease in 3.1 fxm band
strength.
Montmorillonite
Hovolenotn turn)
Salient spectral features in the room temperature smectite spectrum include the
following: l) a deep, asymmetric 1.4 yiim OH + HgO band, 2) a deep asymmetric 1.9 jJ.Tn
absorption, 3) a 2.2 fim Al-OH band, and 4) complete saturation in the region between
~2.75 and 3.2 /xm, which includes both OH and HgO fundamentals. As physically ad-
sorbed water is lost through 160°, both the 1.4 (J.m and 1.9 yum bands decrease in
strength, narrow and become more symmetric. This occurs because the molecules
which contribute to their absorptions are in more well-defined crystalline sites. When
191
the water bands decrease in strength, they are less able to mask other absorptions,
and the apparent 2.2 yLim band increase is due to this effect. At longer wavelengths,
both the 2.9 and 3.1 fim HpO fundamental bands lessen dramatically, eliminating sa-
turation in the 3 /Um region. Loss of tightly held, surface-bound water by 400° contri-
butes to a decrease in band strength at 1.4, 1.9, 2.9, and 3.1 fxm. Very likely the in-
creased slope between 0.6-1.8 /.im is due to cation oxidation by some of this released
water. By 750° structural reorganization begins to occur. The 1.4 /^m band is now both
broad and very weak, suggesting a disruption of sites in which OH is commonly located
as well as its loss from the clay. The miniscule 1.9 ^im feature indicates that little H^O
remains, a conclusion substantiated by the weak 2.9 and 3.1 /Um bands. Incipient
structural reorganization is also indicated by a broadening of the 2.2 /zm Al-OH absorp-
tion.
Paiagonite -' --i— •- ^-'
PAHALA ASH
1
hfl/f n
MavelenQtn Cum}
At room temperature the hygroscopic paiagonite displays the following four
features: 1) a broad Fe electronic transition band centered near 1 fxm, suggesting
the presence of some short range crystalline order, 2) a deep, asymmetric 1.4 /zm HgO
+ OH band, 3) a deep, asymmetric 1.9 /im HgO band on which the resolution of a 1.7 yum
slioulder feature is evidence for extrerae hydration, and 4) complete saturation of the
2.7-4.3 jj,m region. Heating to 120° drives ofl the most loosely bound water molecules,
decreasing the strength of the 1.4 and 1.9 fj,m features and easing saturation between
3.1-4.3 ^im. Lessening these HgO features enables a previously masked 2.2 fira Al-OH
band to be seen. More water, and loosely held OH molecules are driven off by 300°. The
lack of a well defined crystalline network eliminates the distinction of structural OH
ions; hydroxyls are held much less tightly and may be driven off at lower temperatures
than in crystalline clays. Removal of some hydroxyls eliminates the 1.4 jum band and
lessens the 2.75 fj,m band. However, the distinct 2.2 (im Al-OH feature indicates that
tightly bound hydroxyl ions are still held. By 400° loss of these tightly held hydroxyls
begins, indicated by reduction and broadening of the 2.2 ^tm band and continued
reduction of the 2.75 {im band. More HgO is also lost by this temperature, and lessen-
ing 1.9, 2.9, and 3.1 jum bands reflect fiiis change. The trends noted above continue
through 750°.
REFERENCiS: 1) Feierberg. M.A., Lebofsky. L.A., and H.P. Larson (1981) Spectroscopic
evidence for aqueous alteration products on the surfaces of low-albedo asteroids,
Geoch. Cosmoch. Acta 45. 971-981. 2) Pimentel. G.C.. Forney. P.B.. and K.C. Herr (1974)
Evidence about hydrate and solid water in the Martian surface from the 1969 Mariner
infrared spectrometer. JGR 79, No. 11. 1623-1634. 3) Mackenzie. R.C., The Differential
Thermal Investigation of Clays, Mineralogical Society, London, 1957, p. 144-145. 4)
Grim, R.E., aay Mineralogy, 1st ed., McGraw-Hill, New York, 1953. p. 176-177.
ACKNOWLEDGEMENTS: This research was funded by NASA Planetary Geology and MDAP
grant NSG-7590.
192
Studies of the scattering/absorption properties of minerals
Roger N. Clark
U.S. Geological Survey, Mail Stop 964, Box 25046, Federal Center
Denver, CO 80225
An understanding of the scattering and absorption properties of a
planetary regolith is important for understanding remote sensing data from
spacecraft and earth-based telescopes. One way to study these properties
is to measure the reflectance of powdered minerals and mineral mixtures in
the laboratory. This has been and continues to be a major task of many
laboratories. However, the possible combinations of mineral mixtures,
grain sizes, and viewing geometries are virtually infinite. Another
approach to this study is by radiative transfer models. The Hapke reflec-
tance theory (Hapke, 1981) is one such theory that has been verified by
several investigators (e.g. Clark et al., 1986 and references therein).
Using the Hapke theory, the scattering and absorption properties can be
quickly computed as a function of wavelength, grain size, and viewing
geometry for pure minerals as well as mineral mixtures .
Reflectance spectra were computed for water ice and ammonia ice mix-
tures as functions of weight fraction, grain size, and viewing geometry to
simulate possible outer-solar-system satellite surfaces. This exercise has
shown several interesting aspects of scattering and absorption from a par-
ticulate surface that have not been previously realized. As might be
expected, as the grain size of a pure water ice or pure ammonia ice is
increased from very small (< l/^m) , the absorptions increase in depth. How-
ever, at some grain size, depending on the fundamental band strength, the
bands become saturated and their apparent depth decreases. This band
saturation is described in Clark (1981) for water ice.
In a mineral mixture, if the grain sizes are held constant and the
weight fractions are varied, the observed absorption bands change along
with the weight fraction. As the weight fraction increases, the apparent
absorption bands for that species tend to become more apparent (Figure 1) .
However, if the grain sizes are varied and the weight fraction is held con-
stant, the apparent absorption bands still vary (Figure 2)! This is
because the photons are encountering grains according to the projected sur-
face area (of the grain), so either weight -fraction or grain-size changes
will affect the relative surface area encountered. The implications of
this fact are profound for laboratory studies of mineral mixtures. A sim-
ple weight -fraction series is not adequate for deriving a calibration curve
of abundance. The calibration curve depends strongly on the grain sizes.
Because of band saturation, the relative band depths are not the same at
different grain sizes when the grain-size ratios (of minerals in a mixture)
are held constant. The only solutions are either to measure all possible
grain sizes and weight fractions to derive calibration curves, or to use a
radiative transfer model like the Hapke reflectance theory.
As a next step in the understanding of scattering in a particulate
surface, viewing geometry was added to the model calculations. It was
found that viewing geometry had only a small effect on the reflectance
193
levels and observed band depths. Generally, it was found that extreme
viewing geometry ranges (incidence and emission angles near 80°) were simi-
lar to changes in grain size of only factors of two or three .
Everyday experience gives a qualitative understanding of viewing
geometry effects on a multitude of objects; a colored object or a tjrpical
powdered mineral can be held at different orientations and the colors do
not change significantly. However, if two powders are mixed or ground to
smaller grain sizes, the colors can change dramatically, even if the view-
ing geometry is not changed.
In conclusion, reflectance spectra of planetary surfaces are most
affected by the weight fraction and grain sizes of the minerals in the sur-
face. The reflectance can range from 1,0 to about 0.01 by changing the
grain size or weight fraction, a factor of 100. Viewing geometry changes
the reflectance by about 25% or less .
References
Clark, R.N. , Water Frost and Ice: The Near-Infrared Spectral Reflectance
0.65-2.5 fim, J. Geophys . Res., 86, 3087-3096, 1981.
Clark, R.N. , K.S. Kierein, and G.A. Swayze, Experimental Verification of
the Hapke Reflectance Theory 1: Computation of Reflectance as a
Function of Grain Size and Wavelength Based on Optical Constants: J.
Geophys. Res. submitted, 1986.
Hapke, B. , Bidirectional reflectance spectroscopy 1. Theory. J. Geophys.
Res. 86, 3039-3054, 1981.
Figure Captions
Figure 1. The reflectance spectra for mixtures of water and ammonia ice
are shown for a constant grain size. As the weight percent of ammonia
is increased, the ammonia bands (e.g. the 2.2-fim bands) become more
prominent .
Figure 2. The reflectance spectra are shown for 70 -weight -percent water
ice and 30 -weight -percent ammonia ice mixtures with varying grain
sizes. As the grain size of the ammonia ice increases, the ammonia
bands become smaller and the spectrum appears more like that of water
ice.
194
g
l.S
2. 19 2. S
WAVELENGTH Qimi
3.S 3. 5
Figure 1
u .6 -
-| — I — 1 — I — I — I — 1 — I — I — I — |-
2. 2. 5
WAVELENGTH Oi>">
3. 3. 5
Figure 2
MID-INFRARED SPECTROSCOPIC INVESTIGATION ¥-16^ 305 10-37915
3ohn W. Salisbury, U.S. Geological Survey; Louis Walter, Goddard Space Flight
Center, and Norma Vergo, U.S. Geological Survey
This research effort includes; 1) the development of new instrumentation to
permit advanced measurements in the mid-infrared (2.5-25.0 /am) region of the
spectrum; 2) the development of a spectral library of well-characterized mineral
and rock specimens for interpretation of remote sensing data; and 3) cooperative
measurements of the spectral signatures of analogues of materials that may be
present on the surfaces of asteroids, planets or their moons.
New Instrumentation ; Although bidirectional reflectance data in the 3-5 /jm region
of the spectrum can be used directly to predict the spectral behavior of a low
temperature (less than 5C ) remotely sensed surface, emittance begins to dominate
over reflectance at higher temperatures or longer wavelengths. Then, directional
spectral emittance (i.e. in the direction of the observer) dominates. In order to
predict the spectral behavior of remotely sensed surfaces under these conditions, it
is necessary to measure directional hemispherical reflectance (Nicodemus, 1965).
A close approximation of directional hemispherical reflectance can be
determined with an integrating sphere. The inside of the sphere must, however, be
coated with a diffusely reflecting gold surface in order to function properly in the
mid-infrared. We have contracted for the construction of such a sphere, which is
to be delivered in November and will undergo its first tests in December, 1986.
This sphere will provide the first spectral reflectance data of minerals and rocks
from which directional spectral emittance can be confidently predicted.
More detailed, though much more time-consuming, measurements of direc-
tional reflectance can be made with a bidirectional reflectance device with
adjustable angles of incidence and reflectance. Such a device is currently being
designed at Goddard Space Flight Center and is due for completion in April, 1987.
This device should make possible confirmation that the spectral contrast of a
reflectance measurement at zero phase angle will be unaffected by particle size,
as predicted by Salisbury and Eastes (1985). If so, reflectance measurements using
lasers from orbiting spacecraft will avoid the ambiguities inherent in measure-
ments of the spectral emittance of materials of different particle sizes.
Spectral Library ; A library of mineral spectra has been started using a fixed angle
bidirectional reflectance attachment. Relatively pure minerals, generally obtained
from the Smithsonian collection, were crushed and further purified by hand picking.
They were then checked for purity and further characterized by petrographic
microscope (L. Walter), X-ray diffraction (N. Vergo), and electron microprobe (L.
Walter) techniques. Thus, all samples studied have high purity so as to avoid
spectral anomalies, and have been well characterized both mineralogically and
chemically.
So far, more than 100 mineral samples have been processed and approxi-
mately 87 selected for spectral signature measurements (3. Salisbury). For each
sample, the reflectance spectrum of the solid material is recorded, preferably in
different crystallographic orientations, and then reflectance spectra are obtained
of 7^^-250 Aim and 0-7^^ fim size ranges. Finally, a small portion of the 0-7'^AJm size
range is ground to less than 2 /jm and incorporated into a KBr pellet for a
transmittance measurement.
196
The procedure described above yields a variety of spectral data appropriate
for different observational conditions or for surfaces of different particle size
ranges. One important result of these early nneasurements is to show that quite
different spectral features must be measured in remote sensing of coarse and fine
particulate surfaces. This is illustrated in FigurejS 1 and 2, which show reflectance
spectra of olivine (11% forsterite) from 4600 cm" (2.17 jum) to WO cm" (Z^.Oyum).
In both spectra a less than 30 /am particle size range was used, but firmly packed in
Figure 1 and sifted in Figure 2. Packing increases the effective particle size
(Salisbury & Eastes, 1985), resulting in prominent reflectance peaks (reststrahlen
bands) associated with the molecular vibration bands. Of special interest is the
double peak due to the Si-O stretching vibration between 1000 cm" (10 /um) and
SOO cm" (12.5 Aim), because it is in the region of terrestrial atmospheric
transparency. A reflectance spectrum of the same olivine sample sifted into the
sample holder (Fig, 2) shows that this reststrahlen band has greatly decreased in
spectral contrast due to the increased porosity (Salisbury & Eastes, 1985). It also
shows a completely new spectral feature in the form of the broad peak centered
about 700 cm" (14 /jm). This peak is not directly related to the molecular
vibration bands, but is instead associated with transparent behavior due to the low
absorption coefficient between the stretching and bending vibrations at longer
wavelength (Salisbury et al, 1986). We have been able to show, however, that the
wavelength of this transparency peak is just as diagnostic of composition as are
those of the reststrahlen bands. Thus, compositional remote sensing of a fine
particulate surface using the transparency peak can be accomplished despite the
loss of spectral contrast of the reststrahlen bands that make them difficult to
detect (Salisbury and King, in preparation).
Cooperative Efforts ; Several colleasuges have an interest in the mid-infrared
spectral signatures of meteorites, proton bombarded frost residues and even more
exotic materials that may be analogues of materials present on the surfaces of
asteroids, planets or their moons. We encourage the use of our facility to obtain
spectra of these analogues for comparison with telescopic spectral data. To date,
we have had such cooperative efforts with Dale Cruikshank of the Un. of Hawaii;
Tom Jones, a graduate student of 3ohn Lewis' at the Lunar and Planetary
Laboratory of the Un. of Arizona, Tucson; and Trude King of the University of
Hawaii and the USGS, Denver.
References
Nicodemus, F.E., 1965, Directional reflectance and emissivity of an opaque
surfaces Applied Optics, v 4, p. 767-773.
Salisbury, 3.W. and Eastes, J.W., 1985, The effect of particle size and porosity on
spectral contrast in the mid-infrared; Icarus, v 64, p. 586-588.
Salisbury, 3.W., Hapke, Bruce, and Eastes, 3.W., 1986, Usefulness of weak bands in
mid-infrared remote sensing of particulate planetary surfaces: 3ournal of
Geophysical Research (in press).
197
OLIVINE, Foil, PACKED -aoUM POWOER. UABnAOORi K13009
7 Jul 8S lOi S7< 12
bi a
O D
4 600. O 3800. O 3000. O 2200,0 1 BOO. O 1400,0
WAVENUMBERS <CM-1>
lOOO, O BOO, OO
Figure 1. Bidirectional reflectance spectrum of packed olivine powder (less than
30 um).
OUIVJME, FOIL SIFTED -30UM POWDER, UABRAOOR. K13005
7 Jt-il es Qf *Bi IC
Figure 2. Bidirectional reflectance spectrum of sifted olivine powder (less than 30
um).
198
THERMAL-INFRARED SPECTRAL OBSERVATIONS OF
GEOLOGIC MATERIALS IN EMISSION
Philip R. Christensen and Sharon J. Luth, Department of Geology, Arizona State
University, Tempe, AZ 85287.
Thermal-infrared spectroscopy provides a powerful tool for determining the
composition of planetary surface materials. Virtually all silicates, carbonates, sulfates,
phosphates, oxides, and hydroxides have thermal-infrared spectral features associated with
the fundamental vibrational motions of the major ionic groups in the crystal structure. The
vibrational frequency of these motions, and therefore the wavelength of energy absorbed,
varies with both the ionic composition and crystal lattice structure. This variability
provides a direct means of identifying the composition of many geologic materials, and for
interpreting the crystal structure, and therefore the mineralogy of these materials. Because
rocks are composed of mixtures of mineral phases, this technique also permits the
petrology of the rocks and soils exposed at the surface to be determined.
The Thermal Emission Spectrometer (TES) instrument is a thermal-infrared
spectrometer that is currently being developed for use on the Mars Observer mission. To
support this investigation, a series of laboratory measurements of candidate martian
materials has been initiated. These observations are intended to characterize the spectral
properties of geologic materials in emission, and to study a variety of processes and surface
modifications that may influence or alter the spectra of primary rock materials.
Pioneering studies of the thermal-infrared absorption characteristics of minerals
were have been performed (Lyon, 1962, 1964; Farmer, 1974; Hunt and Salisbury, 1974,
1975, 1976). The vast majority of these studies have measured the spectral properties of
materials in either transmission or reflection, whereas remote sensing observations of
planetary surfaces measure the amount of energy emitted. In theory, reflectance
measurements can be related to emission measurement assuming Kirchoff s law, where the
fraction of emitted (e) and reflected (R) energy are related by:
e=l-R
However, this relationship is only strictly valid for measurements of the total hemispherical
reflection and emission from a mat surface, rather that the reflection from a polished surface
as is often measured (Hunt and Vincent, 1968).
In the studies reported here, we have acquired thermal-infrared spectra of materials
in emission. These observations were acquired using the prototype TES brassboard
spectrometer. This instrument is a Fourier transform interferometer, which covers the
spectral range from 7 to 18 }J,m at a spectral resolution of 5.5 cm^l (0.055 |i,m at 10 [im
wavelength). This resolution is approximately a factor of two better than will be achieved
with the Mars Observer instrument, allowing the detailed spectral properties to be
investigated and interpreted. The instrument is controlled, the spectra are acquired, and the
data are processed and analyzed using an IBM PC-XT microcomputer.
To acquire a spectra the sample is heated in a temperature-controlled oven for 4-6
hours to achieve a uniform temperature throughout the sample. At present, only solid rock
samples have been investigated, in order to minimize the effects of temperature gradients
within the samples, which are known to cause uncertainties in the observed spectral
199
properties (Logan and Hunt, 1970; Logan et al., 1975). The samples are removed from the
oven and their spectra acquired within 20 sec to again minimize the effects of temperature
gradients. Spectra are also acquired of a very accurately controlled blackbody reference
surface, and of liquid nitrogen to determine the instrument response and background
emission respectively. This suite of observations permits absolute determination of the
emissivity of the sample, provided that the kinetic temperature of the sample is known.
Because of the difficulty in measuring and interpreting the surface temperature of the
sample, the temperature is determined by fitting a blackbody curve to the observed spectra.
This technique is illustrated in Figure 1. The Michelson mirror position is accurately
determined using a visible light interferometer to count the fringe patterns of a neon light
source. This standard technique permits an absolute wavelength determination to be made
that is accurate and repeatable to approximately 0.01 jim at 10 |J,m. Double-side
interferograms are acquired to permit phase offsets and non-linearities in the beamsplitter
and interferometer to be removed.
Figure la shows the initial spectrum acquired for a quartz crystal, together with the
best-fit blackbody curve used to estimate the sample temperature. Figure lb show the same
data after the observed spectrum has been divided by the blackbody to determine the
emissivity of the crystal. Figure 2 shows a similar set of spectra taken of a basalt sample
returned from the Cima volcanic field in the Mojave desert, California. This sample was
acquired from a flow surface contained within a region imaged by the Thermal Infrared
Mapping Spectrometer (TIMS) airborne instrument (see Kahle and Goetz, 1983). Figure 3
gives the emissivity of a carbonate sample acquired from within the same TEMS scene.
These spectra confirm that thermal emission spectra contain the same absorption
features as have been previously observed in transmission and reflection spectra, and
demonstrate the successful utilization of the TES prototype instrument to obtain relevant
spectra for analysis of rock and mineral composition.
References
Farmer, V.C. (edt.), 1974, The Infrared Spectra of Minerals. Mineralogical Society,
London, 539 pp.
Hunt, G.R., and R.K. Vincent, 1968, The behavior of spectral features in the infrared
emission from particulate surfaces of various grain sizes, J. Geophys. Res.. 73. 6039-
6046.
Hunt, G.R., and J.W. Salisbury, 1974, Mid-infrared spectral behavior of igneous rocks.
Environ. Res. Paper 496-AFCRL-TR-74-0625, 142 pp.
Hunt, G.R., and J.W. Salisbury, 1975, Mid- infrared spectral behavior of sedimentary
rocks, Environ. Res. Paper 510-AFCRL-TR-75-0256, 49 pp.
Hunt, G.R., and J.W. Salisbury, 1976, Mid-infrared spectral behavior of metamorphic
rocks. Environ. Res. Paper 543-AFCRL-TR-76-0003, 67 pp.
Kahle, A.B., and A.F.H. Goetz, 1983, Mineralogic information from a new airborne
thermal infrared multispectral scanner. Science. 222. 24-27.
Logan, L.M. and G.R. Hunt, 1970, Emission spectra of particulate silicates under
simulated lunar conditions, J. Geophys. Res.. 75, 4983-5005.
Logan, L.M., G.R. Hunt, and J.W. Salisbury, 1975, The use of mid-infrared
spectroscopy in remote sensing of space targets, in Infrared and Raman Spectroscop y
of Lunar and Terrestrial Minerals . C. Karr, (ed.). Academic Press.
Lyon, R.J.P., 1962, Evaluation of infrared spectroscopy for compositional analysis of
lunar and planetary soils, Stanford Research Inst. Final Report Contract NASA 49(04).
Lyon, R.J.P., 1964, Evaluation of infrared spectrophotometry for compositional analysis
of lunar and planetary soils: Part II: Rough and powdered surfaces, NASA Contractor
Report CR-100, 172 pp.
200
QUARTZ
QUARTZ
Fig. la
Fig. lb
BASALT
BASALT
Fig. 2a
Fig. 2b
CARBONATE
CARBONATE
Fig. 3a
Fig. 3b
201
Bidirectional Reflectance Properties of Planetary Surface Materials
B. Buratti, W. Smythe, R. Nelson, V. Gharakhani (JPL/Caltech) , B. Hapke (U.
of Pittsburgh)
The JPL spectrogoniometer is capable of measuring the bidirectional
reflectance properties of planetary surface materials for arbitrary viewing
geometries including small phase angles (Smythe et al., 1986a, b). These
measurements depend on the textural characteristics of the regolith, par-
ticularly its packing state, and the albedo of the surface, which dictates
the degree of multiple scattering. In general, porous, dark surfaces should
exhibit the largest opposition surges. Our laboratory measurements on the
JPL spectrogoniometer over the past year have concentrated on separating the
effects of surficial texture and albedo. Figure 1 (Smythe et al.) for
example, which represents a comparison of compacted and fluffy basalt,
demonstrates that if all other factors such as normal reflectance and
composition are equal, less densely packed surfaces exhibit more sharply
peaked phase curves near opposition.
However, dark, porous surfaces are not the only ones with large
opposition effects. Although very dark materials (e.g., carbon black)
exhibit large surges, our measurements of compact barium sulfate show
significantly non-linear increases in intensity, similar to, though somewhat
smaller than the large surges exhibited by very dark materials (e.g. carbon
black) .
The range of materials measured over the past year include charcoal,
several types of basalt, barium sulfate, sulfur, sugar, magnesitom oxide,
several clays, and halon.
Understanding the opposition effect requires a threefold attack:
laboratory measurements, the comparison of these measurements with remote
sensing observations, and the development of a theoretical model to describe
both sets of data. We have extended our comparisons of laboratory measure-
ments to remote sensing data. Figure 2 shows a comparison of lUE observa-
tions of lo with our measurements of fluffy basalt (-90% void space) and a
theoretical model. The UV geometric albedo of lo is comparable to the
normal reflectance of our sample. Our results suggest a fluffy regolith for
lo, similar to that expected from the deposition of material falling from
volcanic plumes. This result contrasts with our finding last year that
Europa most likely has a compact surface with about 25% void space.
A number of technical improvements have been made to the goniometer
during the past year. First, the quartz beam splitter we utilized to
achieve normal incidence and emergence angles introduced unacceptable
secondary reflections into our apparatus. A pellicle beam splitter manufac-
tured by Oriel Corporation was substituted with success -- the secondary re-
flections , disappeared entirely. Secondly, we found that mechanical
stability is a major controlling factor in achieving small phase angles. We
found that sufficiently accurate alignment of the laser, beam splitter,
sample holder, and detector could be attained only by completely substitut-
ing optical XYZ mounts for optical rails and jacks.
202
We incorporated a Mie scattering program into our theoretical model
(Buratti, 1985). The purpose of this approach was to explain the opposition
surges we observed in samples that are not dark and porous. As yet we have
not been able to explain these surges as Mie scattering phenomena.
REFERENCES
Buratti, B. , 1985. "Application of a Radiative Transfer Model to Bright Icy-
Satellites." ICARUS, 61, 208-217.
Smythe, W. , Buratti, B., Nelson, R. , Hapke, B. , Gharakhani, V., 1986a. "Re-
flectance Measurements of Planetary Surface Materials at Small Phase
Angles." B.A.A.S., 19, p. 761.
Smythe, W. , Buratti, B., Nelson, R. , Hapke, B. Gharakhani, V., 1986b. "A
Spectrogoniometer for Measuring Planetary Surface Materials at Small Phase
Angles." Applied Optics, submitted.
This work was sipported by Grant # NAS 7-918.
203
<
o
z
0.4
0.2
10
15
20 25
PHASE ANGLE
30
35
40
45
Figure 1. A comparison of fluffy (~90% void) basalt and dense (-25% void)
basalt showing that fluffy regoliths exhibit larger opposition surges than
dense ones .
T
• lUE
A VOYAGER
O GONIOMETER DATA -
■— SHADOWING /vCiDEL
O.^J
88% VOIDi
21
41 6\ 81
PHASE ANGLE (DEGREES);
io:
121
Figure 2. A comparison of fluffy basalt of about 90% void space with remote
sensing observation of Io and our theoretical model. Data are normalized to
unity at 6 degrees .
204
Atlas of Reflectance Spectra of Terrestrial, Lunar
and Meteor i tic Powders and Frosts from 92 to ISOO nm
Jeffrey Wagner? Bruce Hapke and Eddie Wells
Dept. of Geology 8< Planetary Science? Univ. of Pittsburgh
The reflectance spectra of powdered samples of selected
minerals? meteorites? lunar materials and frosts are
presented as an aid in the interpretation of present and
future remote sensing data of solar system objects. Spectra
obtained in separate wavelength regions have been combined
and normalized? yielding coverage from 92 to 1800 nm.
Spectral features include relfectance maxima in the far UV
region produced by valence-conduction interband transitions?
and reflectance minima in the near UV? visible and near IR
regions? produced by charge transfer and crystal field
transitions. Specific maxima and minima a,re diagnostic of
mineral type and compostion? additionally? the minerals
present in mixtures such as meteorites and lunar samples can
be determined.
205
DECONVOLUTION OF SPECTRA FOE INTIMATE MKTUBES Joha F. Mustard, Carle M. Pieters and
Stephen F. Pratt, Brown University, Providence, RI 02912.
Visible to near-infrared reflectance spectra of macroscopic mixtures have been shown to be linear combinations of
the reflectances of the pure mineral components in the mixture (1). However, for microscopic or intimate mixtures the
mining systematics are in general non-linear (2,3). The systematics may be linearized by conversion of reflectance to
single-scattering aJbedo (w), where the equations which relate reflectance to w depend on the method of data collection
(bihemisphericial, directional hemisphericial, or bi-directional reflectance) (4). Several recently proposed mixing models
may be used to estimate mineral abundances from the reflectance spectra of intimate mixtures. These models are
summarized below and a revised model is presented.
Johnson et al (5) present a semi-empirica] mixing model based on the work of Hapke (4). They used directional-
hemisphericial reflectance data measured in the laboratory and combined pure minerad spectra to fit the spectrum of a
mixture using an iterative least squares approach. Although the flts of endmember spectra to mixttu-e spectra were
within experimental error, the predictions of mineral abundances were 2% to 30% from the actual values. A more recent
approach has been to apply principal components analysis (PCA) to collections of laboratory and lunar spectral data
measured as or, converted U> directional-hemisphericial reflectance (6,7). This technique may be used to determine the
type of mixing (macroscopic vs. intimate), identify the potential endmembers from a suite of spectra and estimate the
relative proportions of endmembers in a sample. The advantages of PCA are the identiflcation of endmembers with
limited a priori knowledge and approximate mixing relations between endmembers. The disadvantages are that it
reqtiires careful interpretation of die principal axis of variation and the quaintitative determination of endmember
abundances becomes complex for mixtures containing more than two components.
We have used a similiar curve fitting approach as Johnson et al (5) with the following exceptions. We use a non-
iterative ( i.e. linear) least squares approach and the data, measured as bi-directional reflectance with incidence and
emergence angles of 30' and 0° were converted to w by the following simplified version of Hapke's (4) equation for bi-
directional reflectance
r(i,e)= w (2.2 -)- H( m )H( m ^)) H( u )= H-2m d)
4( n -^ it ^) 1 + 2 M7
where r is the measured reflectance, i is the incidence angle, e is the emergence angle n =cos(e), u -=cos(i) and y
=(l-w)" . Hapke (4) predicts that for intimate mixtures, the mean single-scattering albedo is related to the single-
scattering albedo of each of the components by
w( X ) = ( X wj( X )MjPidj)/( I MiPjdj) (2)
where M:, P|, and d: are the mass fraction, density and mean particle diameter of each of the i-components in the
mixture. The endmember mineral spectra were combined to fit the measured mixture spectra by a single least squares
inversion of equation (2). The fractional contribution of each endmember is given by the relative geometrical cross
section (f) where ^=(Mj/pjdj)/( T Mjj/pjjd_) for the jth component of an n-component mixture.
This model was testea wiui two mixture series composed of 45-75 um particles: an anorthite-enstatite series (two
minerals of similiar overall albedo) and an olivine-magnetite series (two minerals with strongly different overall albedos).
The quality of the least squares fits as determined by the standard deviation are not greater than 5 x 10' in eJl cases.
The standard deviation is determined by curve shape only and is not sensitive to the eflects of sample preparation which
may result in up to a 2% offset in the measured reflectance curves (5). Shown in Fig. 1 are the observed and least
squares fit spectra for the anorthite-enstatite mixture series.
The computed and actual values for f are shown in Figs. 2 and 3. It is evident from Fig. 2 that, for mixtures of
comparable albedo, equations (1) and (2) can be used to predict the known fractional abtmdances to within the
experimental error. Iliis degree of accuracy is more than sufficient for the purposes of estimating surface mineral
abundances from known or assumed endmembers. The simplified version of Hapke's (4) equation is less satisfactory for
mixtures containing low albedo materials as shown in Fig. 3. The model consistenUy overpredicts the olivine content and
underpredicts the magnetite abundance. The systematic deviation of computed from actual values of f for the mixtures
containing iow albedo minerals suggests that Bither & more rigorous treatment of Hapke's (4) equations is needed or an
empiricial a4jnstment to equation (1) will allow a more accurate solution to be developed.
These data indicate that the simplified version of Hapke's equation for bi-directional reflectance (equation (1)) may
be used to deconvolve reflectance spectra into mineral abundances if appropriate endmembers are known or derived from
&ther techniques. For surfaces that contain a significant component of very low albedo material, a somewhat modified
version of this technique wiU need to be developed. Due to the nature of equation (2) above, this model can also be easily
expanded for use with multicomponent mixtures. Since the abundances are calculated using a non-iterative approach,
tiie application of this method is especially efficient for large spectral data sets, such as those produced by mapping
spectrometers (NIMS,AVIRIS, etc. )
Rsferences: (1) Singer, R.B^ and McCord, T.B. (1978) Proc. Lunar Planet. Sci. Conf. 10th. 1835-1848. (2) Nash, D.B.,
and Conel, J.E. (1974) J. Geophys. Res. fi, 1615-1621. (3) Singer, R.B. (1981) J. Geophys. Res. 86, 7967-7982. (4)
Hapke, B. (1981) J. Geophys. Res. 86, 3039-3054. (5) Johnson, P.E., Smith, M.O., Taylor-George, S., and Adams, J.B.
(1983) J. Geophys. Res. 88, 3557-3561. (6) Smith, M.O., Johnson, F.E., and Adams, J.B. (1985) J. Geophys. Res. 90,
C797-C804. (7) Johnson, P.E., Smith, M.O., and Adams, J.B. (1985) /. Geophys. Res. SO, ,C805-C810.
206
u
o
z
<
I-
o
UJ
_l
lj_
u
a. Ala
t. 00
1.^0
l.b0
t. '10
2.20
2, S
WAVELENGTH IN niCRONS
Figure J. Actual and calculated bi-directional reflectance spectra for the anorthite-enstatite mixture series. The dotted-
curves, labelled AN for anorthite and EN for enstatite, are the endmember spectra used to model the mixture spectra.
The smooth curves are the actual and calculated spectra plotted together. The number located in the 0.9 um. absorption
band corresponds to the % mass of anorthite in the mixture. Each pair of calculated and actual spectra are offset for
clarity and to emphasize the quality of the curve fits. At this scale there is almost no difference between the daculated
and actual spectra.
I
' ■> - « ■ '» I .. , . .. | ( - y^ .- | - n ,-- r'rT -|- T -rr'TY'r-T'T^
smm.
„.lj.J.J.JL.i„
Figure 2. Actual and calculated relative geometricial
cross sections (f-parameters) of anorthite and enstatite
plotted against one another for the mixture series shown
in Fig. 1. The calculated values are obtained from the
curve fitting technique described in the abstract. The
calculated values are generally within l.S% of the actual
lvalues which is adequate for estimating mineral
i^tindances in ^he misctures.
Figure 3. This is Ae same type of plot as in Fig 2 for the
olivine-magnetite mixtiire series. The letters correspond
to the mass fraction ratios where A ^25/75, B= 60/50,
C-75/25, D-90/10, 1=95/5, and the first number is
^e % mass of olivine in the mixtures. The lines between
data points were added to emphasise the systematic
sverprediction @f olivine sontent and underprediction of
magnetitai
207
COHP0SIIIONAL INFOWIATiOtl FOR THE MOON: SOIE CMRACTERISTiCS
©F CURRENT NEAR~iR SPECTRA (TELESCOPIC AND LAWRATORY) - ^leM.
Pfeta^, 0^1. tol^fral Sotei^, Brmn lAilwrslty , Pfwld»», ii 029 1 2
i^tr^uctioii. Dyrli^ Ite ^llo piw«i tte sf^tral reflot^ro dwo^t^lstte of a vwlaty of fynsr
sropte 1^^ m^rrt to tte lata^Mory by Jdtii B. Adras, Itesulte @f 11^ m^rroents mtB
resorted In a ^-te ©f p^5«^ W Adams and Mc^Jsrd Cag, PLX af 1970, 71 . 72, 73). Most ®f mm
fludl^ w^^ dfr^tad toward und^^lmfing the rature ml mnposltlon of lunar »lls, sitm tte y^t
majorlfy of r^note m^suranOTts wb tte'lv^ from tte yppK^mosi s^rfwB a)lls. Itese l^Krstory
fn^ranenti mr@ es^itlslly dl^x^tlni^ with tte clo^ erf tte ^llo pnngrmti. Lato^, In tte Me
I970's, telea^lc lnstromwls w^^ «few!(^ tte! ®l!ow®l hf^ prscfsfw mm--lnft&^ ^^tra,
ranp^d^te In ^jeclral resDlutto to tte ld5(F8tory cteta, to te rtlalnrf fir MlvlAiel lurer ®^@^ 3-20
km In dfmeto* i^fng tel^ix|j^ M tte hfgh-sltftyds Haung Ke@ Obssrysiory ( I ). For tte l@st da^
Mm*- Inf rm^^K)trs f^ small lusw ar^ tev® bem palnsteicfi^ly o^tsinrt by re^arcto^ fran tte
Unfv. of H§walf aid fran Snwn Unlw^lly, SiMrfi taleax^lc spajtr®, with tte 'grsund truth' foumfetlm
ftmn lyrar ^mpte, tew te^ ysed extereivsly by ttese lunar ^lentists t@ &SskWB ® multitude of
pr^Ims In lun^ $c\mm. Summ^te of tte otmpe^ltlmal Infcrmatlon tnl^^mt In tte telest^fc
sp^tra mt be fmM In (1,2,3). Mudt of tte r^r-lnfra^ ^)eetr®! roflectaf^ cMa fretn toth
laterstcry md tsle^x^io me^uromonts tev@ roc^tly teen omplM In oomp^^lo firmats and m
Initial ©HTspa^atfve ^sesOTWt of tte available data m\m elected m^red ^t^ctral p^emete^ tes
bm\ mate. Tte objective Is to develc^ a framewcrk f^ tte syst^natte of lunar nm*-IR s^ro In
ordr to tett^ Int^prete ^}ectra of unknown mat^lals In t^ms of u^ul ampc^ltlci^l Infermatlm.
Spectral Parameters. Eadi sj^ctrum was first classified msor^im to Its gen^^l dwael^. Major
categtr jes Indicated In tte key ara^upmlng Flares i-d sxount for dKJut 2/3 of tte Ata. A variety of
p^met^s positive to mineralogy md alt^ation pr(Mlucts ws^e m^red f^ aadi s^lrum: tmtst ©f
an d)srpllon b^ w&r I pm , stwjgth of tte baral. bml width, IwkI ^mmetry , c^tlnuum slope (near
tte I urn bml), etc. Ttese pertlcul^ properties w^^ ^itosm becmm Important lunar minerals ( high-
ami low-Ca pyroKenes. ©Ilvlr^, glass, plagloolsse) have characteristic absorptloTB ttet affect tte
strer^h and nature of these pw^met^s In dlagrastlc ways ( 1 ,2,3), Example of tte relation tetweai
^m@ of these ^)eclral parmate^s for tte Vm collections are shown In figure s-d. Eadi ^ectrum was
sl^ ^slgied a relative quality factor from 1 to 5; only quality I and 2 have teen used In th^e f l^ires.
felesoopk Data. Tte current data tese lnclu(te slm^ 3S0 Imllvldual ^)ectra for about 2 1 © areas 3
to 1 5 km In dimtBtBr i IfM^s^Klent measurements ire required for mf^t areas). If ctota scatta^ amcmg
r^ea^diers Is lnclu(^, tte lun^ collection could te ^tmi 1 /3 larg^. A\Vm^ sane attempt tm Ismi
made In recent years to ^sure most Important geologic units it@ Included In tte ^IMIm, tte survival
of data ^xiulsltjai progrOTs depencfe on productive scl^x^. and tte cMa rallectlcn Is nes^^'lly bl^
toward are^ studied for specific tc^te (mar^ soils, Oc^jernlcus «wl Its rays, hlgt>l«J oat^^s, etc.).
Latioratory Data. Tte Adams' ^lectra include almost 420 spectra of "ISO lunar sample. About
701 are lunar soils representing all Apollo sites, as well as Luna 20 and 24. Although tte s^le of a
IdKratory sample Is sSxnA I ifi tlm^ smaller than tte a^ of teles^lc mae^rOT^ts, tte two ty^ of
data are direct^ comparable for tte relatively well-mixed lunar soils. However, for lunar rocks and
minerals, and especially for breoclas, tte collection of sample spectra Is Incomplete due to tte difficulty
of obtaining or idaitifying mat^ials from limited sanples ttel are representative of s rode, a site, cr
even tte moon as a wtele. Not all lunar rock types Identified are represented by sample spectra
Comparisons. Tte Urn collecticms are not expected to te entirely similar for tte reeswis mmUtxM
above ^^rning ^mplet^t^s. A few well-known lunar ^lectral properti^ &r@ evident In teth, mA
m tte systmatiD variation In w^rwm^ €«Dposltion (detect by ter^ poslti^ afU strer^h) betw^i
Ihg highlands CLo~€s) and inir® CHI-€e). Ni^. howev^, that en eddilional gabbroc CHI-Cs px)
iMiponmt mt be dst^rt C te^^ bend ewtw) in mm hn^lmd ffstart efFi.iC3}. Aean Kwnpte of
208
raipsitfoBl fnfewi«, m^l^ the fllresy miplm (a^ m 74220, 15401 «d a synltetfc Apllo
1 2 flte) wlilch wstwialtally mMMl wm- hm^ thm tte melii lr«* ^ rrts rt alls In Flfl. &
The pr^lretfc mmtlfr^ mrt^fel et Arlstarchis, Sulpfcfys Mlm, md Slfw» ^tuym ^hfbft
stelllo'ly braJ bai&, froplylri Fs-^li^ gl^ *^fis (4). OlMm, wh « Ihsl at ^s^nirais'
witrel p^ im tm slmlla^ pn^M^lies. tat «i oftw be dlstfowl^iad freiii §]m by ottw-
dw@Dl«^fette mh m dlbeii. Tte regta mrted W ot Fig ^ «rf d Irelydg ^lls from mymi^ m-ll
bssslls 0f the w^to^n lu^la IteM r^tsml mlh for which ro smplm wist («teJ o-os on Ftp. a
afrf c) but whlA clwly witalre substsitlal ollvire ir o!^5).
AclsgwIsdjmsBls^ This r«®m^ h sopporltd isy MASA grMi NA®ii^-28. Rersresess. I) HeCt^ ti al..
ML */m m no. Bl f, plC»83.- 2) FfelOT fta> tPSCTO*. p2825; 3) Pi§t®r3.l%5, ^#k5. S$ophys., to
Fts« 4) ii^iiiLlLJSS. icstm,6f, p45l| S) Pitiiyssial. I^TO. j®?, #5, n©. B7, p39l3,
i i §
_, — I — , — I — , — p
TELESCOPIC DATA 1.2
X
+
o
m
t
• Soils
Highland Soils
Mar© Craters
Highland Craters
Highland nountains
Mantling Material
ll,.og|
STRENGTH OF
1 Un BAND
I
as
aW
i
I
air
STRENGTH OF 1 UN BAND
<
CD
lu^ar" s.Anf=
T ^
S.AnPLE 1. 2
00
_L
I
asr
I
I
"a^a-
CD
Z
Ld
»-
Z
Ld ?
U
• 1 . 1 . , .
TELESCOPIC DATA 1.2
1 . -
d_
s
s
-
8 o
•J
IS
ate ®.se ass
STRENGTH OF 1 UM BAND
:!?
9
® 8 B
air
0.^0
STRENGTH OF l' UN BAND
209
C^MSiamM Siraiigl^&YQf CctsM Biatrial tem Hear- Infrared Spectra
C. M. Pieters {Bro'wi UiiiTrersiiy)
The -walls mi4 electa swro-uading large imf^t craters 40-100 km in diameter
expose tlie first few kilometers of crystsl material at the target site jiBt t>elovthe
surface. Central peaks of sucli craters expo^ material derive<3 from greater
-tepths ("4 - 10 km) |^. Grieve and Head, LPS; 14 1. High spatial resolution
measwements of the composition of material ^sociated with such craters can t>e
used to probe the subsurface criist. An earth-based telescopic program to acquire
near-infrared spectra of freshly exposed liMar material nov contains dataltsr 1?
large impact craters -with centra peaks. Noritic, gabbroic, aiorthwitic and
troctolitic rock types cmi be distingtmshed for areas 'lithin these large craters
from charaeteiistic absorptions in indlviduai spectra of their -roills and central
pe&ks. Horites dominate the ^per itmar cn^st vMle the deeper crostal lones also
contain significant amounts of gabbros on4 anorthosites. Examples of the
compositional stratigraphy of the lunar highland crmx observed at specific
craters include (subsurface crust - deeper crust):
Aristarchw: gabbro 4- troctolitic gabbro - gabbro [Ltjcey et a! L^ 16 1
Bullialdus: gabbro - norite
Copernicm: norite - troctolite fPieters et al, JP£ M 12,393 1
Eratosthenes: norite - gabbro
Langreni^: norite throughout
Lansberg: aorite - norite (^th clinopyroxene)
Theophilus: norite - anorthosite
Tycho: gabbro throughout
Vhen the complexity of planetary crusts are compared, the Moon is often
considered a relatively simple and homogeneous end-member in terms of crtastal
formation processes as ^^11 as the mixing that must have occmred diMng the
extensive early bombardment history. These data for material asscx^ated wth
large craters, hoirever, indicate that not only is the lunar crust highly
heter^eneom across the nearsi<te, but that the compositional stratigraphy of the
lunar crt^t is also nonwiiform. If such diversity occiars for even this small
planet, crmtal complexity should be expected for other planetary bodies to be
studied. For maximum aMipositlonal information iKing advanced sen^srs on
future mimons C&alileo, MO, LGO) me«wfement strategy shoMd incltide high
spatial and spectral reMluiion 4aia in aid aromid lo-ge impact craters.
210
PiffiUMMARY EESULTS OF SPECTRAL REFELECTANCE OTODIES OF TYCHO
CKATEE
B.R. HAWKE, C.R. COOMBS, P.G. LUCEY. and J.F. BELL, Planetary Geosciences
Division, Hawaii Institute for Geophysics, University of Hawaii, 2525 Correa
Road. Honolulu. HI 96B22: R. JAUMANN and G. NEUKUM, DFVLR. 8031 Wess-
ling, ¥. Germany; C.M. PIETERS, Department of Geological Sciences, Brown
University, Providence, RI, 02912.
INTM}DUCTIOM
Spectral studies of lunar craters and their ejecta deposits provide valuable
information about the composition of the lunar crust as weE as insights into
the impact cratering process. Materials from a variety of depths have been
excavated by these impacts and deposited in a systematic manner around the
craters. Hence, inform.ation concerning vertical and lateral changes in the
composition of the lunar crust can often be obtained by careful analysis of
spectral data of these ejecta deposits. Results of multispectral imaging stu-
dies of Tycho, Aristilus, Lalande, and Kepler were published by Hawke et at.
(1979). Lucey et al. (1986) more recently have published the results of a
spectral reflectance study of Aristarchus crater and its deposits. This paper
presents the preliminary analysis and interpretation of near-infrared spectra
obtained for both the interior and exterior deposits associated with Tycho
crater. Specifically, our objectives were: 1) to determine the composition
and stratigraphy of the highland crust in the Tycho target site. 2) to deter-
mine the likely composition of the Tycho primary ejecta which may be
present in ray deposits, 3) to investigate the nature of spectral units defined
in previous studies, 4) to further investigate the nature and origin of both the
bright and dark haloes around the rim crest, and 5) to compare the composi-
tions determined for the various Tycho units with those of Aristarchus crater
as well as "iypicaT highlands deposits.
METHOD
Near infra-red spectra (0.6 - 2.5 /im)for small (~2 km) areas in and around
Tycho were obtained at the Mauna Kea 2.24-m telescope using the Planetary
Geosciences Division indium antimonide spectrometer. Extinction correc-
tions were made using the techniques described by McCord and Clark (1979).
Analyses of absorption bands and continuum slopes were made using the
211
methods presented by McCord et al. (1981). Spectra were collected for the
following areas: 1) the southwest crater wall, 2) the northern rim, 3) the
southwest floor, 4) the eastern floor, 5) the central peak, 6) the northeast
ejecta blanket, and 7) the dark halo north of the crater. The residual absorp-
tion spectra (after continuum removal) are shown in Figure 1.
KSSULTS ANP DISCU^ION
The spectra obtained for areas on the interior of Tycho exhibit similar spec-
tral features. These include relatively strong l^m absorption bands whose
minima are centered between 0.97 and 0.99 ^m and shallow to intermediate
continuum slopes. The spectra generally exhibit indications of a 1.3 fxm
feature consistent with the presence of Fe -bearing plagioclase feldspar.
The strong 1 fjxn absorption features indicate a dom.inant high-Ca clinopyrox-
ene component. These spectra are all Type G as defined by Pieters (1986).
The material exposed on the interior of Tycho is a gabbroic assemblage with
compositions ranging between gabbro and anorthositic gabbro.
The spectrum that we obtained for the Tycho central peak is slightly different
from that presented by Srarekar and Pieters (1986). Their spectrum exhi-
bits a shallower continuum slope, a deeper 1 /i,m band centered at a slightly
longer wavelength (0.99 /zm vs. 0.98 fJ-m), and a well-defined 1.3 fjjxi absorption
feature. Their spectrum was collected for a larger area of the Tycho central
peak and may better represent the character of peak as a whole than does
the one we present in Figure 1.
Spectra were also obtained for Tycho ejecta deposits. Hawke et al. (1979) ,
noted that the Tycho rim crest is surrounded by two spectrally distinct
haloes. The inner, bright halo is narrow, discontinuous, and poorly developed
south of the crater. The outer, dark halo is m.ore extensive and continuous
and extends in places to almost one crater diameter from the rim crest. One
spectrum was collected for an area just north of the Tycho rim crest. This
area appears to be typical of the inner, bright halo. The spectrum is almost
identical to those obtained for interior units and a similar composition is im-
plied.
The spectrum obtained for a portion of the dark halo north of Tycho is very
different from those of the interior units. This spectrum exhibits a wide, rela-
tively shallow absorption feature centered at 1.01 /xm, a 1.3 ^m absorption,
and a relatively steep continuum slope. This spectrum is very similar to that
212
presented by Smrekar and Pieters (1986) for a portion of the dark halo
southwest of Tycho. They interpreted their spectrum as indicating the pres-
ence of pyroxene, Fe-bearing feldspar, and a significant component of Fe-
bearing impact melt glass. We agree with this interpretation and suggest that
the dark halo north of the crater is composed of a similar assemblage.
The spectra of spots inside of Tycho
show, in the context of the highlands
thus far investigated, the greatest
similarity to certain spectra obtained
for features in the Aristarchus region
(i.e.. Class 1) of Lucey et ai. (1986).^
However, the suite of spectra obtained
for Tycho exhibits a diflerent trend
than that of the Aristarchus spectra in
terms of band center versus width.
Aristarchus spectra show a strong
correlation between band center and
width which Lucey et al. (1986) attri-
buted to a variation in olivine content
in the highland units in the region.
This is not the case at Tycho. Ap-
parently, olivine is absent or present
in only minor amounts in the Tycho re-
gion.
HGUREl:
Spectra of locations of Tycho crater
interior and exterior deposits, with
straight line continua removed. (A)
Dark halo, north (B) Northeast ejecta
(C) Southwest wall (D) Southwest floor
(E) East floor (F) North rim (G) Tycho
central peak.
wavelength (um)
REraRENCE?:
1) B.R. Hawke et al. (1979). Conf. sm the Lunar Highlands Crust. 50; 2) P.
Lucey, et al., (1986), PLP5C 16. in press; 3) T. McCord and R. Clark (1979),
Pub. Astron. Soc. Pac. 91. 571: 4) T. McCord. et al., (1981) JGR. 86. 10883; 5)
C. Pieters (1986), Rev. Geophysics, in press; 6) S. Smrekar and C. Pieters
(1986), Icarus, in press.
213
Freilniinary Results of Geologic and Remote ^nsiog Studies of Rima Mozart
C.R. COOMBS and B.R. HAWKE, Planetary Geosciences Division, Hawaii Institute
of Geophysics, University of Hawaii, 2525 Correa Rd., Honolulu, HI 96822.
The nature and origin of lunar sinuous rilles have long been
the subject of major controversy. Lunar sinuous rilles typically
occur on mare surfaces, though a few cross highland terrains.
Suggested origins for sinuous rilles vary widely: 1) erosion by
nuees ardentes (Cameron, 1964), 2) fluvial erosion (Urey, 1967;
Gilvarry. 1968; Lingenfelter et al, 1968; Schubert et at., 1970), 3)
tectonic features of tensional origin (Quaide, 1965) 4) formation
as lava tubes and channels during lava emplacement (Kuiper et
at., 1966, Greeley, 1971, Oberbeck et aL, 1969) or, during the
draining of a lava lake (Howard et at., 1972), 5) incision of chan-
nels by thermal erosion and /or turbulent flow through the chan-
nels (Carr, 1974; Hulme. 1973).
It is very unlikely that nuees ardentes would form the
smooth sinuous channels on the lunar surface. On the Earth
they form hummocky channels and dunes, not smooth curvy
channels. Though a morphologic similarity to terrestrial fluvial
features has been noted (Urey, 1967; Gilvarry, 1968, Lingenfelter
et al., 1968), a fluvial origin for lunar rilles has long been
discounted due to the anhydrous nature of the returned lunar
samples (e.g. Taylor, 1975). The similarity of certain lunar rilles
to terrestrial channels of fluvial origin merely suggests an origin
by fluid flow, though water is not necessarily the erosive agent
(Oberbeck et at., 1971). Terrestrial lava tubes and channels
often originate at vent craters or depressions associated with
regional tectonic features such as faults, fissures, or fracture
systems that often form concentric to calderas (Greeley, 1971).
Similarly, many lunar sinuous rilles are located in areas of struc-
tural weakness and often tend to follow pre-existing structural
trends. However, the presence of elongate "head" craters,
levees, and the fact that lunar sinuous rilles may extend for
large distances without offset or mirror images on either side
helps negate an origin soley by extensional tectonics.
Lunar sinuous rilles are also similar to terrestrial lava chan-
nels and collapsed lava tubes in that they commonly originate at
irregular shaped "head" craters, trend down-slope, are discon-
tinuous in areas (tube formation and /or collapse?), taper out at
distal ends, and may form distributaries. Thermal erosion and
turbulent flow may help .explain the extent of lunar sinuous
rilles, but it can not explain the formation of meanders that are
commonly associated with lunar rilles.
214
In order to better understand the processes responsible for
the formation of lunar sinuous rilles, we have conducted a study
of Rima Mozart using a variety of geologic, photographic, and
remote sensing data. The apparent source of this rille is located
in a highlands unit of known composition (i.e., KREEP basalt;
Hawke and Head, 1978; Spudis, 1978; Spudis and Hawke, 1986)
and it was hypothesized that thermal and mechanical erosion
played an important role in the formation of Rima Mozart.
Excellent photographic, topographic, multispectral, and radar
data exist for this rille. The purpose of this paper is to present
the preliminary results of an analysis of this data.
Rima Mozart is a 40 km-long lunar sinuous rille located near
the SE rim of Imbrium basin (25°2l"N, 359°03"W). It is situated
about 100 km southwest of the Apollo 15 landing site. It is 250 -
500 m deep and is incised into the volcanically derived KREEP
basalts of the Apennine Bench Formation and dark mare basalts
surrounding the rille (Figure l).
Photographic data indicates the presence of two volcanic
source vents for Rima Mozart. Now girdled by dark mare-type
material, these vents appear to have formed in the underlying
Apennine Bench Formation (Figure l). The primary source vent,
Kathleen, is an elongate crater (5x3 km) in the Apennine Bench
Formation. From Kathleen, Rima Mozart follows a dominant
NW/SE structural trend (Swann, 1986) until terminating 30 km to
the east. About 10 km from the termination of Rima Mozart the
second elongate (2x1 km) source vent, Ann, is joined to the
main channel by a secondary rille. Evidence of spatter sur-
rounding Ann strongly supports the volcanic origin of this
feature. At the distal end of Rima Mozart, the rille terminates at
Michael, a possible "sink" crater. The crater Michael may be
connected to the NE-trending Rima Bradley, to the southeast,
through a conduit and /or underground plumbing system marked
by a collapse feature , Patricia. This possible extension of Rima
Mozart joins Rima Bradley at a NW/SE angle, reflecting the dom-
inant structural trend of the region.
¥e suggest that Rima Mozart, like many other lunar sinuous
rilles, is most likely formed by a combination of events. Rima
Mozart does follow a pre-existing, dominant, NW/SE structural
trend suggesting the influence of structural features on the rille,
however, the tectonic influence is not the sole source for the for-
mation of the rille, as suggested by the presence of two volcanic
source vents and spatter present around Ann., We suggest that
the rille formation began with an explosive eruption at Kathleen
which later calmed down to a pulsating, high volume, low-
viscosity lava flow. The rapid effusion rate of the magma as well
as its high temperature and turbid nature helped carve the
215
sinuous rille into the fractured and structurally weak Apennine
Bench Formation underneath. Similar eruptions and subsequent
flows were also created at Ann and joined to the main channel by
a NE-trending secondary rille. Rapid and turbulent lava flows
continued to form Rima Mozart through a distributary system of
channels and/or tubes that paralleled the underlying, pre-
existing structure of the Apennine Bench Formation until reach-
ing the terminus at Michael.
"^.
-SL.
"^^
'4v
X
o
Ui
ffl
LU
RIMA MOZART ''YAvi^^^^l^^
a.
<
^^'-V \V; • PATRICI
{
km
MARE
'jr.
MGURE 1: A sketch map of the Rima Mozart region showing the locations of
the two surce vents, Kathleen and Ann, and the "sink" crater, Michael.
fCSS: (1) Cameron, W.L. (1964), IGR, fi^ 2423; (2) Urey, H.C, (1967), Nnt.nrR. 21fi. 1094; (3)
Gilvany, J.J. (196B), w«tiirp^ 2JB, 336; (4) Lingenfelter, R.E., Peale, S.J. and Schubert, G. (196B),
S^i.'-nf^p IfijL, 266; (5) Schubert, G., Lingenfelter, R.E. and Peale. S.J. (1970), Bex. of Gfophys, and
SpanpPTiyg, R, 199; (6) Quaide, W. (1965). Icflrus,4. 374; (7) Kuiper. G.P, Strom, R.G.. LePoole, R.S.
(1966). lEL Lah. T^nh. B^-pt , 32-800; (8) Greeley. R. (1971), Sfiiennp, 122, 722; (9) Oberbeck. V.R.,
Quaide. W.L. and Greeley. R. (1969). Wn^pTr, n^nlngy, j, 75; (10) Howard, KA., Head, J.W. and
Swann, G.A. (1972), ^r,r- 3tH T.imar SoiPnPf. TnTif., Vn1.i>, 1; (ll),Carr, M.H. (1974), Icarus, 22, 1;
(12) Hulme, G. (1973), Modem Gf^nlogy, 4, 107; (13) Taylor. S.R. (1975), Lunar Science- A £oat
Apniln ViPw, 1; (14) Swann, G.A. (1986). T.PSr JOH, 855; (15) Hawke and Head (1978). Proe.. flth
r.iiTi»r PlaTiPt ScL rr.Tif 3285; (16) Spudis (1976). Eroc fith Lunar Elaufit- ScL C.rmf., 3379 ; (17)
Spudis and Hawke (1966), Prnc Apo11nl5CoTif,, 105.
216
PHOTOMETRIC PROPERTIES OF LUNAR TERRAINS DERIVED FROM HAPKE'S EQUATION.
Paul Helfenstein and Joseph Veverka, Center for Radiophysics and Space Research,
Cornell University, Ithaca, NY 14853-6801.
Hapke's bidirectional reflectance equations [1,2,3] provide the most rigor-
ous available description of photometric behavior in terms of physically mean-
ingful parameters. In the newest version [3], six model parameters characterize
photometric properties: The efficiency with which average particles scatter and
absorb light is described by single scattering albedo, w. The opposition effect,
a surge in brightness observed in particulate surfaces near zero phase (4,5,6,7,
8,9,10) is characterized by two parameters; h, which relates its angular width
to the combined effects of regolith compaction and particle size distribution,
and S(0), the contribution of singly Scattered Tight primarily by first-surface
reflection at zero phase. The particle phase function is a Legendre polynomial
with first and second order coefficients b and c, respectively. Finally, 5, is
an average topographic slope angle providing a measure of sub-resolution scale
macroscopic roughness.
The complexity of Hapke's equation and comparatively large number of model
parameters makes the unique solution of meaningful parameter values difficult.
Reliable methods for simultaneously solving all six parameters from remotely-
sensed data needs development and testing. The primary objective of this study
is to derive, Hapke parameters for the lunar surface from both disk-integrated
and disk-resolved photometric data. The Moon is the only extraterrestrial body
for which samples have been returned to Earth for laboratory analysis, for which
extensive manned and unmanned geological field observations have been conducted,
and for which corresponding regional photometric measurements have been obtained
over adequate ranges of illumination and viewing geometry. The derivation of
physically and geologically meaningful parameters from the existing lunar
photometric data set is an important test of the usefulness of Hapke's equation
in remote sensing applications where surface-based ground truth measurements are
unavailable. Hapke's parameters for the Moon also provide a foundation for the
comparison of parameters derived from other planetary surfaces.
Visual disk-integrated phase curve data were obtained by renormalizing and
combining the measurements of Russell [11,12], Rougier [13], and Shorthill et
al . [14] to the precise photoelectric measurements of Lane and Irvine [15],
which covered a slightly smaller range of phase angles. The combined data were
subdivided and averaged in phase angle bins having the following intervals:
data in the range of 0°- 10°were divided into 2,5° bins, data in the range of
10° - 30° were divided into 5° bins, and data in the range of 30° to 150° were
divided in 10° bins. Lunar disk-resolved reflectance data were a subset of the
extensive measurements tabulated by Shorthill et al. [14] who divided terrains
into three groups on the basis of their normal albedoes. An. We have preserved
their classification system and used data for thirty-two dark lunar terrains
(0.06< An <0.09), twenty-five average terrains (0.10< kn <0.12), and thirty
bright terrains (0.13< An <0.16) in their catalogue. We combined all similar
terrain classes into three groups and renormalized all brightness measurements
to group-mean albedoes. Average brightnesses, incidence angles (i), and emission
angles (e) were determined at each phase angle ( a ) in each group.
Hapke's equation was fit to the disk-integrated phase curves and disk-resol-
ved data for dark, average, and bright terrain classes using an iterative, non-
linear least-squares algorithm described by Helfenstein [16J. Parameters were
initially determined from the disk-integrated data, and the result was applied
as a first-guess to the iterative solution of parameters for individual terrain
classes. Table 1 lists our numerical results. The visual phase curve computed
from the best-fit is presented with the averaged disk-integrated data in Figure
1. Figure 2 shows plots of the disk-resolved data normalized to corresponding
brightnesses predicted from the disk-integrated solution under the same illumin-
ation and viewing geometries. The curves in Figure 2 represent the similarly
normalized brightnesses predicted from disk-resolved solutions for all points
such that i=e= a/2.
Disk-integral photometric properties of the Moon and Mercury at visual
wavelengths are known to be remarkably similar [17]. Our value of w, S(0), b,
and 9 are nearly identical to those found by Veverka et al. [18] for Mercury.
Our value of h (0.07) falls in between Veverka ^ jil .' s Mercurian value (0.08)
and the 0.05 value derived from lunar farside data [19] by Hapke [3]. Disk-
integrated and disk-resolved solutions are mutually consistent. The value of
each disk-integrated parameter lies between end-member dark and bright terrain
values, and does not differ strongly from the average albedo terrain value.
Systematic trends in disk-resolved parameters can be identified in Table 2.
Not surprisingly, values for single scattering albedo (w) of the dominantly
anorthositic average and bright terrains are significantly larger than the value
for the basaltic dark terrains (mare). Decreasing values of b for dark through
217
Table 1: Hapks Parasaters for Lunac Tast&ins
Disk-Integcatsd
w
0.21
Disk-Resolved Dark 0.12
Disk-Resolved Average 0.25
Disk-Resolved Bright 0.33
h S(0)
0.07 0.71
0.12 0.51
0.06 0.78
0.05 1.03
b c ff
0.29 0.39 20.0*
0.41 0.10 8.1*
0.33 0.37 20. 6»
0.26 0.4S 24.0"
PHASE ANCU
figure 1 daft): prsdictsd llghtcucve and averaged disk-integrated data
(one sigiaa error bars), noraalized to geoaetric albedo at zero phase.
Figure 2 (bottom): Averaged disk-resolved data noraalized to predicted
brightness from dtsli-integratad solution, a) dark terrains, J>) average
terrains, c) bright terrains. Curves represent predicted normalized
brightnesses for hypothetical points on the surface such that i-e-aC/2.
c.
1 1 1 1 1 1
-^- . . 1 . . . . 1 . . ■
PHASE ANCU
PtUiSS Max
bright terrains in(3icate that dark r
tering than those in average and bri
decrease from dark through bright te
the opposition effect is greatest fo
mare soils have lower surface porosi
characteristic regolith particle-siz
of the lunar surface are known to be
difference is most likely due to a g
in lunar highlands regoliths [23].
Bq " S( 0)/w{l+b+c} , for dark terrain
terrains. This appears to be a cons
a larger fraction of singly scattere
surface reflection, S(0). The avera
is significantly lower than_that of
of 1 for bright terrain to 9 for dar
to subcentimeter scales and compares
radar RMS slopes of highland (^-8®)
egoliths are, as expected, more backscat-
ght lunar terrains. Values of h likewise
rrains indicating that the angular width of
r mare. This would suggest either that
ties, or that they have a slightly different
e distribution. Since porosities over most
remarkably similar (/^45%) [20,21,22], the
reater abundance of fine grains {<20^im)
The total amplitude of the opposition surge,
s is larger than for average and bright
eguence of the fact that in opaque particles
d light at zero phase comes from first
ge macroscopic roughness (9) of dark terrain
average and bright terrains. The ratio of
k terrain is 2.5. This value corresponds
well to the ratio of 2.0 derived from
and mare (^-4") at 13 cm wavelengths [24],
Acknowledgement: This research was supported by NASA grant NSG-7506.
REFERENCES: [1] Hapke, B. (1981) J^ 86 , 3039-3054. [2] Hapke, B.(19B4) Icarus
59' 41-59. [3] Hapke, B. (1986) Icarus 67, 264-280. [4] Seeliger, H.(1887)
Abhandl . Baver . Akad . Wiss . Math .- Natur . 16 , 405-516. [5] Seeliger, H. (1895)
AiitiaQdi. Baver. Akad . Wi??. Math. Hatvr. JJ., 1-72. [6] Hapke, B. (1963) 2S&. §1.
4571-4576. [7] Hapke, B. (1966) Astrophvs . J. n . , 333-339. [8] Bobrov, M.
(1970) In Surfaces and Interiors of Planets and Satellites. Acad. Press, NY.
[9] Gehrels, T. , et al,. , (1964) Astron . J. 69, 826-852. [10] Irvine, JGR 71.
2931-2937. [11,12] Russell, H. (1916) Astrophvs . J. 43, 103 & 173. [13] Rougier,
M. (1933) Ann. Obs . Strasbourg _2, 205-339, [14] sKorthill, R., et ai.. , (1969)
NASA Contr. Rep. CR-1429, 405 pp. [15] Lane, A. and W. Irvine (1973) Astrophvs .
J. 78, 267-277. [16] Helfenstein, P. (1986) Ph.D. Thesis, Brown Univ. , Providence,
RI, 414 pp. [17] Hapke, B. (1977) Physics of Earth and Planet . Interiors 15,
264-274. [18] Whitaker, E. (1979) Icarus 40. 406-417. [19] Veverka, J. et al.,
(1986) in preparation. [20] Houston, W., et al.(1972) Proc . LPSC III . 3255-
3263. [21] Carrier, W. , et al.(1972) Proc . LPSC III . 3223-3234, [22] Carrier,
W" et al.(1974) Proc . LPSC V, 2361-2364. [23] McKay, D,, H. il. ( 1974) Proc .
LPSC V, 887-906. [24] Tyler, G, (1979) Icarus 37. 29-45.
218
Bidirectional Reflectance Spectroscopy
k. The Extinction Coefficient and the Oppostion Effect
Bruce Hapke
Dept. of Geology &c Planetary Science, Univ, of Pittsburgh
Icarus &7, aijif-SBO (1986)
The extinction coefficient and the opposition effect in
a particulate medium are discussed. Simple analytic
expressions that describe these quantities are rigorously
derived using a few physically realistic mathematical
approximations. The particles of the medium may have a
distribution of sizes and the particle density is allowed to
vary with depth. The xpression for the extinction
coefficient is valid for both large and small porosities and
is more accurate than the one commonly used. The opposition
effect arises from the hiding of extinction shadows and
occurs even if the particles are transparent. The angular
half-width of the opposition peak is shown to be equal to the
ratio of the average particle radius to extinction length at
unit slant path optical depth in the medium, and depends on
both the filling factor <ratio of bulk to solid density) F
and the particle size distribution.
To illustrate the theory it is fitted to observations of
the moon, an asteroid and a satellite of Uranus; Europa is
also discussed. For the moon a value of F=0.^1 is derived,
in good aggrement with data on Apollo soils. For Oberon, the
width of the oppositon effect peak gives F=0.10, which is
similar to values for terrestrial frosts and snow. Thus, the
narrow opposition effects of the Uranian satellites do not
require any unusual particles or microstructures on their
surfaces. More photometric observations of Europa are
needed .
219
On the Sputter Alteration of Regoliths of
Outer Solar System Bodies
Bruce Hapke
Icarus 66, 270-279 (1986)
This paper discusses several processes that are expected
to occur when the porous regoliths on bodies without
atmospheres in the outer solar system are subjected to
energetic ion bombardment. The conclusions reached in many
of the papers involving sputtering published in the planetary
literature are qualitatively or quantitatively incorrect
because effects of soil porosity have been neglected. It is
shown theoretically and experimentally that porosity reduces
the effecitve sputtering yield of a soil by more than an
order of magnitude. Between 90 and 97X of the sputtered
atoms are trapped withn the regolith? where they are
factionated by differential desorption. Experiments indicate
that more volatile species have higher desorption
probabilities. This process is the most importanbt way in
which alteration of chemical and optical properties occurs
when a regolith is sputtered. When a basic silicate soil is
irradiated these effects lead to sputter-deposited films
enriched in metallic iron? while O, Na and K are
preferentially lost. The Na and K are present in the
atmosphere above the sputtered silicate in quantities much
greater than their abundances in the regolith. Icy regoliths
of SOg. should be enriched in elemental S and/or SejO. This
prediction is supported by the probable identification of SsO
and polysulfur oxide bands in the IR spectra of H-sputtered
SOss reported by Moore <198'4-, Icarus 59, 11A-). When porous
mixtures of water? ammonia and methane frosts are sputtered?
the loss of H and surface reactions of C> N and in the
deposits should produce complex hydrocarbons and
carbohydrates? some of which may be quite dark. Such
reactions may have played a role in the formation of the
matrix material of carbonaceous chondrites prior to
agglomeration.
220
Charged Particle Modification of Surfaces In
The Outer Solar System
R»E. Johnson, Univ, of Virginia
Voyager reflectance spectra data have indicated clear leading/trailing
differences in the albedo of the icy Galilean and Saturian satellites.
For the Galilean satellites, these have been analyzed by Nelson, et
al. and, more recently, by McEwen. They have described the longitudinal
dependence of this data and attempted to interpret this in terms of
plasma and meteorite modification of the surface. Primary attention
has been paid to Europa at which the leading/trailing differences are
the largest.
Recently we have reanalyzed this data extracting the single grain
(particle) albedo, w, and constructing the Espat-function, W=(l-w)/w
from this. Because w is near unity, W-^ Z^D where -a is the absorption
coefficient and D the grain size. In doing so we find a direct comparison
to the longitudinal plasma bombardment flux for the first time (see
figure) . This occurs primarily in the UV and is probably due to an
asorption associated with implanted S, as the UV band of Voyager overlaps
the lUE data of Lane et al. We also can now unravel the relative importance
of grain size effects and implant impurity effects.
Work supported by NASA grant NAGW-186
Clarke, R.N., et al. Icarus 56 (1983) 233
Lowe A.L. et al. Nature 292 (1981) 38.
McEwen, A.S. Icarus (in press) (1986).
Nelson, M. et al. Icarus 65 (1986) 129
221
1.4
^
.20 -
Smoothed values of Espat Function
vs. cosine of logltude from apex
of motion
Equltorlal Elux of sulfur ions
bombarding Europa vs. longitude
lOt cmpmisQU of photohetric scams produced by the minhaert and hapke
FUNCTIONS. Damon P„ Simonelli and Joseph Veverka, Cornell University.
Experience has shown that the empirical Minnaert function is a very
useful approximation to real photometric behavior near opposition (phase
angle a = 0°)^ but that in general it cannot accurately model photometric
scans across the face of even a homogeneous planet at higher phase angles
(see for example Goguen, 1981). Given recent work on fitting the rigorous
Hapke photometric function to Voyager data for lo (Simonelli and Veverka,
1986), we can test to what degree the Minnaert function breaks down in the
case of Ionian materials by comparing photometric scans produced by the two
approaches.
At phase angles a= 2, \Q, 30, 60, and 90° ^ we have computed scans of
the reflectance along the photometric equator (photometric latitude Y= 0°)
and mirror meridian (photometric longitude 00= a/2) that would be expected
for a homogeneous planet whose surface obeys Hapke' s law (Hapke, 1981,
1984). We use values of the Hapke parameters uoj h, g, and derived for
lo by Simonelli and Veverka (1986) in both the Voyager narrow-angle camera
violet filter (\ « 0.42 pm) and orange filter (\ « 0,59 pm). Each calcu-
lated Hapke scan is compared with the corresponding scan predicted by
Minnaert's law for various values of the Minnaert limb-darkening parameter
k. For a Minnaert scan at a particular k, we arbitrarily choose the value
of the reflectance parameter Bq so that the Minnaert and Hapke scans coin-
cide at the so-called "specular point," the point where the photometric
equator and mirror meridian intersect (¥ =0°, u= a/2). The violet-filter
photometric scans that result from this process are shown in Figs, 1 and 2;
orange-filter results are qualitatively similar and are not displayed.
VIOLET FILTER PHOTOMETRIC EQUATOR
050
040-
030-
020
0.(0 J — I
70 -50 -30 -(0 (0 30 50 TO
O.SOi
-TO -50 -30 -10 (0 30 SO TO
20-
0=60°
1 1 1 1— — T 1 1 p-
a«90'
Figure 1, Scans at
various phase angles
illustrating how the
reflectance I/F varies s
along the photometric
equator. The solid
curve shown at each
phase angle represents
a scan predicted by the
Hapke function, using
values of the Hapke parameters
derived by Simonelli and
Veverka (1986) from disk-
integrated observations of
lo in the Voyager violet
filter: Sq = 0.68, h = 0.24,
g = -0.14, and 9 = 25°.
The dashed curves shown
at each phase angle are
scans predicted by the
Minnaert function for the
various indicated values of the limb-darkening parameter k; for
dashed curve, the value of Bq has been chosen so that the curve
with the Hapke scan at the specular point (see text).
-TO -so -30 -TO 10 30 SO TO SO
-90 -70 -50 -30 -10 10 30 50 TO tO 30 30 40 50 60 TO 80 90
Photometric Longitude (degrees)
each
coincides
223
It is apparent that the Minnaert law does not match "exactly" Hapke
scans at any of the phase angles shown, even a= 2°. Yet our plots demon-
strate that at least in the case of lo, the Minnaert law is a useful empir-
ical tool. The Minnaert scans deviate most strongly from "real" (Hapke)
behavior close to the limb of our hypothetical planet, i.e., near photo-
metric latitude ±90° or photometric longitude +90°, However, as distances
along the projected disk of a planet visible to an observer go as the sine
of the photometric angles, when compiling photometric information most of
the useful data come from areas within « 60° of the sub-observer point.
Figure 1 shows that along the photometric equator, the Minnaert description
doesn't break down severely until u = e > 70°, which corresponds to x >
0,94, where x is a linear scale from x = at the center of the projected
disk to X = 1 at the limb, and e is the emission angle. Thus the Minnaert
function breaks down most noticeably at geometries which ordinarily
contribute least to observational data sets.
We also note that in the case of lo, the Minnaert description is use-
ful not only at small phase angles, but throughout the range of a consider-
ed here. Figs. 1 and 2 suggest that the Minnaert function is as good an
approximation at a = 90° as at a = 2°; i.e., it seems able to match the
Hapke scans to about the same level of approximation at all phase angles.
VIOLET FILTER MIRROR MERIDIAN
30
Figure 2. Same as
Fig. 1, but for
scans along the
mi rror meridian.
0=2"
-T 1 1 1 T"
a MO*
a»30»
30
0=60°
a =90"
, — I — , — p.
R^tometric Latitude {degrMs}
There is only one respect in which the Minnaert approximation worsens
at high a: it appears that slightly different values of k are needed to
describe scans across the face of a planet in different directions.
Specifically, as the phase angle increases, the value of k that best
matches behavior along the photometric equator is more likely to differ
224
from the k that results along the mirror meridian. For example, if we com-
pare the a = 60° scans for the violet filter in Figs. 1 and 2, we find k «
0,9 along the photometric equators but k « 0,8 along the mirror meridian.
This variation in k with scan azimuth has already been discussed by Goguen
(1981) in another context. It is noteworthy that for lo the effect is not
serious for moderate phase angles (a < 60°),
In summary, our work with lo data indicates that the empirical
Minnaert function, while not a perfect model of real photometric behavior,
does provide a very useful parameterization of limb darkening at phase
angles out to 90°, and is especially useful near opposition (cf,, McEwen
and Soderblom, 1984; Clancy and Danielson, 1981), Those who work with the
Minnaert law, however, must bear in mind the major limitation inherent in
this empirical function: Minnaert parameters are unspecified functions of
the phase angle. While Bo(a) for a specific material typically drops with
increasing a, and k(a) generally increases toward higher phase angles (a
trend seen in Figs. 1 and 2, as well as in Harris, 1961; Goguen, 1981; and
McEwen and Soderblom, 1984), the determination of a material's Bq and k at
one a provides no direct information as to the values of these parameters
at other phase angles.
This research was supported by NASA Grant NSG 7156,
References
Clancy, R. T,, and G, E. Danielson (1981), J, Geophys, Res . 86, 8627-8634.
Goguen, J, D, (1981), Ph,D. dissertation, Cornell University, Ithaca, N,Y.
Hapke, B, (1981). J. Geophys, Res , 86, 3039-3054.
Hapke, B. (1984), Icarus 59, 41-59,
Harris, D, L, (19611^ iF The Solar System III. Planets and Satellites
(G. P. Kuiper and B. M, Middlehurst, Eds.), pp. 272-342. Univ. of
Chicago Press, Chicago,
McEwen, A, S., and L. A. Soderblom (1984), NASA Technical Memorandum
86246, pp. 261-262.
Simonelli, D., and J. Veverka (1986). Phase curves of materials on lo:
Interpretation in Terms of Hapke's Function. Icarus , in press.
225
SULFUR- OXYGEN PROCESSES ON 10
Robert M. Nelson, William D. Smythe
Jet Propulsion Laboratory, Pasadena, CA 91109
After many years of research efforts dedicated to interpreting the spec-
tral geometric albedo of Jupiter's satellite lo, the only chemical specie which is
generally agreed to exist on lo's surface is, condensed sulfur dioxide (Smythe et
al. , 1979). Although the evidence for the presence of elemental sulfur is strong,
there are several important astronomical observations and laboratory studies which
constrain or question the presence of elemental sulfur in widespread areal abund-
ance on the surface. In addition, the presence of sulfur dioxide on lo's surface
raises questions regarding the presence of environmentally produced daughter
products (if any) which also might be expected on the surface. This laboratory
work is directed to understanding sulfur/oxygen processes in the lo environment.
Laboratory studies of irradiated sulfur dioxide frost have found that
sulfur trioxide should be formed as a consequence of the irradiation process. We
have measured the spectral reflectance of solid sulfur trioxide in the laboratory
and we f ind . that it has strong absorption features at 3.37 and 3.70 nm. These
features are not present in the spectral geometric albedo of lo. We interpret this
as an indication that sulfur trioxide may exist in such limited abundance that it
is undetectable in disk averaged spectrophotometry. We suggest that the Near-
Infrared Mapping Spectrometer on the Galileo spacecraft should be able to identify
condensed sulfur trioxide on lo particularly in regions bordering the sulfur
dioxide deposits (Nelson and Smythe, 1986).
The presence of elemental sulfur on lo's surface has been questioned on
several grounds most notably the suggested production process (quenched molten
sulfur extrusions) and the effect of radiation (most notably x-rays) on some of
the allotropes . We have produced mixtures of sulfur allotropes in the laboratory
by quenching molten sulfur and we find that the spectra indicate the presence of
certain red- colored allotropes which are preserved upon quenching. We find that
the color of the sulfur glass produced is redder when the temperature of the
original melt is higher (Nelson and Smythe, 1985). This is consistent with the
suggestion that lo's spectral geometric albedo can be partly explained by the
presence of quenched sulfur glasses.
Preliminary investigations of the effect of x radiation on elemental
sulfur indicates that at low temperatures (-100 deg K.), a red- colored allotropes.
This would be consistent with the presence of elemental sulfur on lo. Further
work in this area remains for a follow- on investigation.
REFERENCES
Nelson, R. M. and W. D. Smythe, (1985). Jupiter's satellite lo: Surface
composition based on spectral reflectance of quenched sulfur glasses. EOS ,
Tr. Am. Geophys . U. , 66, p. 947.
Nelson, R. M. and W. D. Smythe, (1986). Spectral reflectance of sulfur trioxide:
Implications for Jupiter's satellite lo. Icarus , 66, pp. 181-187.
Nelson, R. M. (1983). Color of irradiated sulfur at low temperature:
Implications for lo. Bull . Am . Astron. Soc. , 15, p. 851.
Smythe, W. D., R. M. Nelson, D. B. Nash, (1979). Spectral evidence for sulfur
dioxide frost or adsorbate on lo's surface. Nature , 280 , p. 766.
226
SOLID SULFUR IN VACUUM: SUBLIMATION EFFECTS ON SURFACE MICRO -
TEXTURE, COLOR, AND SPECTRAL REFLECTANCE, AND APPLICATIONS TO PLANETARY
SURFACES; D. B. Nash, Jet Propulsion Laboratory, California Institute of
Technology, Pasadena, CA
A new form of sulfur that is white at room temperature, and very
fluffy in texture, has been found in laboratory experiments on the effects
of vacuum sublimation (evaporation) on solid sulfur. This work is an
outgrowth of proton sputtering experiments on sulfur directed toward under-
standing Jovian magnetospheric effects on the surface of lo (Nash, 1985).
Fluffy white sulfur is formed on the surface of solid yellow, tan, or
brown sulfur melt freezes in vacuum by differential (fractional) evapora-
tion of two or more sulfur molecular species present in the original
sulfur: Sg ring sulfur is thought to be the dominant sublimation (evapora-
tion) phase lost to the vacuum sink, and polymeric chain sulfur (S ) the
dominant residual phase that remains in place forming the residual fluffy
surface layer.
The microscopic structure of the fluffy sulfur layer determined from
scanning electron microscope (SEM) images is skeletal with a lacy-like
fabric and with filamentary components having a scale size of <1 um. It
appears from its relatively organized structure to be dominated by
molecular forces rather than adhesive forces between particulate grains
such as that described for water/clay sublimate residues by Saunders et al.
(1985). In the SEM pictures it resembles cellular sulfur material
described by Tuinstra (1967) resulting from dissolving melt-freeze sulfur
in liquid CS2 solvent and obtaining an insoluble polymeric residue. The
residual subliming surface becomes very porous (up to about 98% void space)
and has a bulk thickness of 0.5 mm after 800 hours at 2 x 10" torr ( 3 x
10 "•'-° atm.)
The color of the evaporating surface changes from the original --
which at room temperature (_300 K) may be yellow, tan, or brown depending
on the pre-freeze melt temperature (393-713 K) - - to white with a faint
greenish-yellow tint. For fresh sulfur melt- freeze samples evaporating at
_300 K in _10" torr vacuum there occur significant spectral, color, and
albedo effects in as little as 10 hours, becoming uniformly white within
300 hours, and progressive changes in spectral details continue for at
least 1200 hours. Once removed from vacuum conditions the material is
stable in form and color with time (for at least several months).
The reflectance spectrum of the original sulfur surface in the UV/VIS
(0.35 to 0.70 um) range is greatly modified by formation of the fluffy
surface layer: the blue absorption band- edge and shoulder move 0.05 to
0.06 um toward shorter wavelengths resulting in a permanent increase in
reflectivity near 0.42 - 0.46 um by as much as 400% or more, and the UV
reflectivity below 0.40 um is reduced to about 1/3 its original level to as
low as 2%. The new band-edge and shoulder positions are temperature
sensitive, as in unmodified sulfur, shifting to shorter wavelengths with
decreasing temperature, and returning to their pre-cooled wavelength with
temperature recovery; but with fluffy white sulfur the temperature- induced
excursion range of the absorption edge is almost entirely within the violet
and UV whereas in ordinary sulfur it is mostly in the green and blue
wavelengths .
227
The sublimation (evaporation) rate of sulfur from fresh frozen sulfur
melt at initial exposure to high vacuum (__10' torr) is on the order of 2 x
10 S cm" sec' at _300 K (in agreement with previous measurements by
Bradley, 1951), increasing steeply with temperature, decreasing with higher
vacuum pressure, and decreasing with vacuum exposure time reaching an
equilibrium flux of about 1-4 x 10 S cm" sec" after _1200 hours.
This remarkable form of lacy, fluffy, white sulfur should exist in
large quantity on the surface of Jupiter's satellite lo, especially in
hotspot regions if there is solid free sulfur there that has solidified
from a melt. Its color and spectra will indicate relative crystallization
age on a scale of days to months and/or surface temperature distribution
history. It suggests that for a sulfur flow surface there could be a class
of surface hotspots that have a white-is-hot relationship, an inverse
variant of, but not in conflict with, the albedo -vs- temperature relation
demonstrated in the lo data by McEwen et al. (1985). The concepts to be
developed from this work on fluffy sulfur are expected to be helpful in
understanding properties of lo's surface such as composition, adsorbtivity,
thermal inertia, polarization, photometry (with solar phase angle), and
post-eclipse brightening. The flux of subliming sulfur from hotspots on lo
could be a copious and continuous source supplying the lo sulfur torus.
Material with structure similar to that of fluffy sulfur could exist
on surfaces of other small (airless) planetary bodies, including comets,
that have multiple surface constituents of differing volatility.
References :
Bradley, R. S. (1951). The rate of evaporation and vapour pressure of
rhombic sulfur. Proc. Roy. Soc. , A205 , 553-563.
McEwen, A. S., D. L. Matson, T. V. Johnson, L. A. Soderblom (1985).
Volcanic hotspots on lo: Correlation with low-albedo calderas.
Submitted to J^ Geophys . Res .
Nash, D. B. (1985). Proton irradiation of sulfur: Sputtering yields,
spectral reflectance changes, and applications to lo. Bull. Am.
Astron. Soc. , 17, 921.
Saunders, R. S., F. A. Fanale, T. J. Parker, J. B. Stephens (1985).
Properties of filamentary sublimation residues from dispersions of
clay in ice. Submitted to Icarus .
Tuinstra, F. (1967). Structural a psects of allotrop y of s ulfu r a nd th e
°£ll^£ ^iZ^i£S£ ^lH^nts^- Uitgerverij Waltman, Delft (The
Netherlands). 110 p.
228
A Preliminary Analysis of the Mariner 10
Color Ratio Map of Mercury
Barry Rava and Bruce Hapke
Dept. of Geology & Planetary Science, Univ. of Pittsburgh
A preliminary geological analysis of the Mariner 10
orange/UV color ratio map of Mercury is given, assuming a
basaltic crust. Certain errors in the map are pointed out.
The relationships between color and terrain are distinctly
non-lunar. Rays and ejecta are bluer than average on
Mercury? whereas they are redder on the moon. This fact,
along with the lack of the ferrous band in Mercury's
spectrral reflectance and smaller albedo contrasts? implies
that the crust is low in Fe and Ti. There is no correlation
between color boundaries and the smooth plains on Mercury, in
contrast the strong correlation between color and maria-
highlands contacts on the moon. The smooth plains are not
Mercurian analogs of lunar maria, and a lunar-type of second
wave of melting did not occur. Ambiguous correlations
between color and topography indicate that older, redder
materials underlie younger, bluer rocks in many places on the
planet, implying that the last stages of volcanism involved
low-Fe lavas covering higher-Fe rocks. There is some
evidence of late Fe-rich pyroclastic activity.
229
CHAPTER 7
RADAR APPLICATIONS
A COMPILATION SYSTEM FOR VENUS RADAR MISSION (MAGELLAN)
Sherman S, C. Wu, Francis J, Schafer, and Annie-Elpis Howingtonj
U»S. Geological Survey, Flagstaff, Arizona 86001
A synthetic aperture radar (SAR) compilation system has been developed for
extraction of topographic information of Venus from stereoradar imagery to be
obtained from the Magellan mission. The system has been developed for an AS-
11AM analytical stereoplotter (Wu et al., 1986). During fiscal year 1986,
extensive tests were made on this compilation software by using stereo-images
from various radar systems, both spacecraft and airborne (Table 1). Maps were
compiled and the precision of planimetry and contour measurement was
evaluated. Digital data of some models were also collected for processing
orthophoto or perspective views by using the original radar images.
From the Seasat radar images (Wu et al,, 1986), a planimetric map was compiled
from the New Orleans model and a contour map from the Los Angeles model.
Because of strong geometries, i, e., large base-to-height ratios (0.99 and
0.85), 0.9-m and 1.2-m repeatability elevation measurements were obtained from
these two model s.
A contour map of Mt. Shasta (Fig. 1) was compiled on the AS-llAM analytical
stereoplotter from SIR-B stereoradar images (same-side mode). Because the
average residual of elevation measurements in this area is about 8 m, the
contour interval could have been as small as 25 m. Residuals of elevation
measurements vary, however, depending on the combination of look-angles of the
pair of stereo-images.
Three stereomodels of the SIR-B images have been compared. Their respective
average elevation residuals are 7.3 m, 8.7 m, and 33.9 m, and their respective
combinations of look angles are 29.5° and 60.1°, 29,5° and 53.5°, and 53.5°
and 60.1°.
The two models of stereoradar images taken by the NASA 102A airborne radar
system cover part of Mt. St. Helens. Both models have a repeatability
elevation measurement of about 5m. A contour map was compiled from each
model, but the geometry of the model having an opposite-side mode (Fig. 2) is
stronger than the one having a same-side mode.
Also tested with the SAR compilation software were two stereomodels of radar
images of the vicinity of Spitsbergen taken by the airborne ERIM X-C-L radar
system. The models cover the same ground features, but one model is in
ground-range geometry and the other is in slant-range geometry. Geometries of
both models are compatible. The model in ground-range geometry was used to
compile a contour map at a scale of 1:25,000 with a contour interval of 10 m.
As noted above, stereomodeling from radar images has been proven feasible and
thus the mathematical concept of the compilation software is the correct
approach. All stereomodels tested were, in general, free of parallax and
suitable bases for map compilation. During testing, the software was enhanced
and modified to obtain more flexibility. Development, of the radargrammetry
will be continued to improve measurement precision and map compilation for
multiple models. Feasibility studies of map compilation from digital radar
data will also be considered.
233
TABLE 1. RADAR IMAGES USED FOR TESTING THE SAR COMPILATION SYSTEM
Vehicle
Spacecraft
Aircraft
Mission/Radar
System
Seasat
SIR-B
NASA 102A
ERIM X-C-L
Site
New Orleans
Los Angeles
Mt, Shasta
Mt. St. Helens
Spitsbergen
Geometry*
S.S,
0,S.
S,S.
O.S,
S.S.
S.S.
Range**
S.R.
S.R,
G,R.
G.R,
G.R.
S.R. & G.R.
Remarks: * S.S. --Same-side nrode; O.S, — Opposite-side mode
** S.R, --Slant-range geometry; G.R, --Ground-range geometry
Reference
Wu, S. S. C. , Schafer, F. J., and Howington, A. E. , 1986, Radargrammetry for
the Venus Radar Mapper Mission: Report of Planetary Geology and
Geophysics Program-1985, NASA Tech, Memo. 88383, 570-573,
234
Figure 1. Topographic contour map of Mt. Shasta. Scale 1:200,000.
interval 100 m; 50-m intermediate contours in northern part.
Contour
235
Figure 2. Topographic contour map of part of Mt. St. Helens compiled on the
AS-llAM analytical stereopl otter. Scale 1:41,700; contour interval
25 m.
236
New Veiy High. Kesoliitioii Eadar -Sudies of tt.e Msson.
Peter J. Mouginis-Mark and Bruce Campbell. Planetary Geosciences, Hawaii
Institute of Geophysics, University of Hawaii, Honolulu, Hawaii 96822
As part of an effort to further understand the geologic utility of radar studies of the
terrestrial planets, we are currently collaborating mth investigators at NEROC Haystack
Observatory, MIT and the Jet Propulsion Laboratory in the analysis of existing 3.8 and 70
cm radar images of the moon, and in the acquisition of new data for selected lunar tar-
gets. The intent is to obtain multi-poleu-ization radar images at resolutions approaching 75
meters (3.8 cm wavelength) and 400 meters (70 cm wavelength) for the Apollo landing
sites (thereby exploiting available ground-truth) or regions covered by the metric camera
and geochemical experiments onboard the command modules of Apollos 15, 16 and 17.
Similar studies were conducted by Moore and Zisk (1973) immediately following the Apollo
Program, but at much lower spatial resolution and without the complete phase history of
the radar echo being recorded.
As of Fall 1986, the acquisition of the new data is well underway. Three attempts have
so far been made by S. Zisk at Haystack to obtain new images at 3.8 cm wavelength, and
preliminary measurements of the leading edge of the moon and the Apollo 17 landing site
show that geologically-useful images are likely to be produced by early 1987. 70 cm data
have also been obtained by Thompson for Copernicus Crater. Currently, improvements
still have to be made to the pointing of the radar antenna, but once the lunar emphemeris
is better knovm and radar tracking has been improved it should be possible to obtain
radar images measuring a few thousand square kilometers at the desired sub-kilometer
resolution.
In the meantime, we are renewing studies of the 3.8 cm lunar maps produced by Zisk
et ai. (1974) and 70 cm data produced by Thompson (1986). These data were collected in
both like- and cross-polarizations, and in the case of the 70 cm data, permit the phase
records to be used to assess the scattering properties of the surface in a similar manner
to the interpretation of terrestrial quad-polarized radar images (Thompson et ai., 1986;
Zebker et al., 1986). In particular, we are comparing the distribution of surface units on
the moon that show a mism.atch between the surface slope implied by like- and cross-
polarized scattering data, based on the scattering models of Evans and Hagfors (1964),
and Hagfors and Evans (1968). The unusual values of cross-to-like polarization observed
for volcanic flows in western Mare Imbrium (Schaber et al., 1975) have been observed in
both of our existing data sets and, together with the Apollo landing sites and young
craters Copernicus and Tycho, form the basis for targeting of new radar acquisitions far
geologically-interesting areas on the moon over the coming months.
REFERENCES: Evans. J.V. and T. Hagfors (1964). On the interpretation of radar reflections
from the moon. Icarus, vol. 3, p. 151-160. Hagfors, T. and J.V. Evans (1968). Radar stu-
dies of the moon. Chapter 5 in: Radar Astronomy, (J.V. Evans and T. Hagfors, Eds.),
McGraw-Hill, N.Y. Moore, H.J. and S.H., Zisk (1973). Calibration of radar data from Apollo
17 and other mission results, ^ollo 17 Prelim. Set Rpt., NASASP-330, p, 33-10 to 33-17.
Schaber, G.G., T.W. Thompson and S.H. Zisk (1975). Lava flows in Mare Imbrium: An evalua-
tion of anomalously low Earth-based radar reflectivity. The Moon, vol. 13, p. 395-423.
Thompson, T.W. (1986). High resolution lunar radar map at 70-cm wavelength. Submitted
to Earth, Moon and Planets. Thompson, T.W., H.A. Zebker and J.J. Van Zyl (1986). Lunar
radar polarimetry. Trans. Amer. Geophys. Union, in press. Zebker, H.A., J.J. Van Zyl and
D.N. Held (1986). Imaging radar polarimetry from wave synthesis. /. Geophys. Res., in
press. Zisk, S.H., G.H. Pettengill and G.W. Catuna (1974). High-resolution radar maps of
the lunar surface at 3.8 cm wavelength. The Moon, vol. 10, p. 17-50.
237
HIGH RESOLUTION RADAR MAP OF THE MOON
T* Ws Thompson s Jet Propulsion Laboratory , Pasadena, CA 91109
Previous radar mappings of the Moon at 70 cm wavelength in the late
1960s by Thompson (1974) have been replaced with a new set of obser-
vations conducted between 1981 and 1984 using the 430 MHz radar at the
Arecibo Observatory » Puerto Rico. Radar resolution was reduced to 2-5
km radar cell -size and a "beam-sweep", limb-to-limb calibration was
conducted. Advances in computer technology provided the principal means
of improving lunar radar mapping at this wavelength. These new lunar
radar maps are described in greater detail by Thompson (1986)»
The antenna beamwidth for the 430 MHz radar at Arecibo is only 10 arc-
minutes, about one-third of the angular width of the lunar disk when
viewed from Earth, Thus, some 18 separate beam positios were needed to
map the entire disk. Twelve of these were placed in an "outer ring"
surrounding the center of the disk. To tie the eighteen separate beam
positions together, a limb-to-limb, beam swing calibration was con-
ducted. Here the antenna beam was slowly swept across the lunar disk at
a rate of 2 arc -minutes per minute. This was repeated north and south
of the apparent equator on several different days. This calibration was
conducted at somewhat coarser resolution than that for the individual
beam positions described above.
Radar observations of the Moon were conducted by transmitting pulses
from the main antenna and receiving echoes at an auxiliary antenna
located some 11 km NNE of the main antenna. Circular polarization was
used to obviate the adverse effects of Faraday rotation in Earth's
ionosphere. Post-observation data reduction used the delay -Doppler
techniques described by Thompson (1978), Radar echoes from the Moon are
separated into time-delay (range) bins and Doppler-frequency bins which
provides a two-dimensional separation and eventual mapping to lunar
latitude and longitude. Radar echo strengths are also normalized in the
mapping by removing predictable variations. This processing removes
background noise levels, accounts for antenna gain and scattering area
differences across the beam, divides by an average scattering law, and
adjusts final map values so their averages on a 5° x 5° squares agree
with the limb-to-limb, beam-sweep calibration described above. The
final map is a square array of pixels separated by 0,1° in latitude and
longitude. Two 1800 by 1800 pixel arrays represent the lunar earths ide
hemisphere (+90° in latitude and longitude) in the two radar polari-
zations. These data have been shipped to the Planetary Data Centers.
The data from these observations is shown in Figure 1, which shows an
orthographic projection of the new radar data. The weakest scattering
differences in these displays show scattering differences on the order
of ten to twenty percent. The largest scattering differences in Figure
1 are those places which saturate as totally white areas have radar
echoes ten or more times stronger than the average. Most echo devia-
tions tend to be stronger than the average (whiter in the photographs of
Figure 1).
238
References:
T»W* Thompson, 1974, Atlas of Lunar Radar Maps at 70-cm Wavelength, The
Moon , 10 . 51 -85.
T»W* Thompson, 1979, A Review of Earth-based Radar Mapping of the Moon,
The Moon and Planets , 179-198*
T,W* Thompson, 1986, High Resolution Lunar Radar Map at 70-cm
Wavelength, accepted for publication in The Earth, Moon, and Planets»
Figure U New high resolution radar map of the moon at 70 cm wave-
length* North is at the top and grid lines are every 15° in
latitude and longitude^, Thus, photograph shows "polarized"
echoes; the expected polarization from a plane mirror
reflection.
239
LANDFORM IDENTIFICATION - LUNAR RADAR IMAGES
Henry J. Moorej U.S» Geological Survey, Menlo Park, CA 94025
T, W. Thompson, Jet Propulsion Laboratory, Pasadena, CA 91109
Three sets of polarized radar-echo images of the Moon are
being examined to establish the relation between radar
resolution and landf orm-identif ication resolution [1-4], The
wavelengths, radar resolutions (cell sizes), and approximate
number of real or apparent landforms for the sets are as
follows 5
Set # Wavelength Cell sizes Number of landforms
1 3.8 cm [5] 1-2 km 1,553
2 70 cm [6] 2.5-5 km 1,594
(high resolution)
3 70 cm [7] 10-20 km 983
(low resolution)
The results of the study should be valuable to those
planning to acquire or interpret radar images of the Earth or
other planetary bodies.
After comparison with lunar maps and photographs, real and
apparent landforms on the radar images are grouped into one of
seven classes [1-4]: (1) resolved and clearly identified; (2)
resolved and would probably be correctly identified; (3)
resolved, but interpretation is uncertain; (4) detected, but
elements are not resolved; (5) not detected; (6) array of
landforms is resolved, but interpretation of the array is
uncertain; and (7) radar portrays a fictitious landform.
Data recorded for each real or apparent landform for each
set of images includes the following: (1) a name, (2)
selenographic coordinates, (3) diameter and relief obtained
from lunar maps and photographs, (4) the class, (5) diameter
measured on the radar image, (6) background terrain, (7)
geologic age, and (8) the Lunar Aeronautical Chart number. A
computer program sorts and orders the data by diameter or
relief, computes the percent of each class in frequency bins
of 100, and computes the geometric mean of the landform
diameter or relief for each frequency bin. Calculations are
made in frequency steps of 10.
Current results show strong relations between radar
resolution and diameter or relief of landforms that are
240
clearly identified and those that would probably be correctly
identified (class 1 + class 2), as shown in table 1« Current
results are not depicted; they are similar, but not identical,
the those in previous abstracts [e,g. 1, figure 1]^
Table 1. Percentage of resolved and identified landforms
portrayed in lunar radar images.
Percentage
class 1+2
landforms
of
Mean diameters of landforms (km)
corresponding to the indicated percentage,
3.8 cm 70 cm high 70 cm low
10
5
12
35
20
6
16
56
30
7
25
—
40
9
35
—
50
11
44
—
60
16
—
__
70
20
. —
--
80
35
—_
__
Percentage of
class 1+2
landforms
10
20
30
40
50
60
70
80
Mean relief of landforms (km)
corresponding to the indicated percentage.
3 . 8 cm 70 cm high 70 cm low
0.5
0.6
0.8
1,1
1.4
2.0
2.2
2.6
1.0
1.7
2.6
2.8
3.1
2.7
Lan
reliefs
with in
Landf or
and rel
radar c
landf or
cm, 70
respect
by the
fictiti
dforms are no
, but the per
creasing mean
ms are simply
iefs. Ambigu
onstitute up
ms at various
cm high resol
ively. Only
radar images
OU8 (class 7)
t detected (class 5) at all diameters and
centage of undetected landforms decreases
diameter and increasing mean relief,
detected (class 4) at most mean diameters
ous arrays (class 6) portrayed by the
to about 16, 22, and 15 percent of the
diameters and relief values for the 3.8
ution, and 70 cm low resolution images,
a few percent of the landforms portrayed
at various diameters and relief values are
When data acquisition is complete, the data will be
analyzed as functions of angle of incidence (lunar-scattering
function) [6], background terrain, and geologic age [7],
Preliminary comparisons of the actual observed crater
241
size-frequency distributions with those obtained from the
radar images show increasing departures with increasing
resolution. For the 3»8 cm radar images, the cumulative
frequency of craters greater than 22.6 km across agree to
within 21 percent of the actual cumulative frequency, and the
population indices (slope of the distribution) are similar and
near -2, Here, the crater frequencies from the radar image
lie below the observed ones. Comparisons for the other radar
images are less satisfactory at this time.
References
1] Moore, H, J. and Thompson, T. W. , 1986, Landform
identification in lunar radar images: Lunar and Planet,
Science XVII, p, 565-566.
2] Moore, H, J, and Thompson, T, W, , 1986, Lunar radar
noore, h, j, ana xnompson, r, w,, lyoo. Lunar radar
images - landform identifications Repts, Planet. Geol. and
Geophys, Prog, — 1985, NASA TM 88383, p. 550-552.
3] Thompson, T, W, and Moore, H, J,, 1985, Landform
identification in radar images: Lunar and Planet, Science
XVI, p. 860-861,
4] Moore, H, J, and Thompson, T, W, , 1985, Landform
identification on radar images: Repts. Planet. Geo!
Geol.
Thompson, T. W» , unpublished data.
Thompson, T, W., 1974, Atlas of lunar radar maps at 70 cm
wavelength: The Moon, v. 10, p. 51-85.
8] Hagfors, T,, 1970, Remote probing of the Moon by infrared
and microwave emissions and by radar: Radio Science, v, 5,
p. 189-227,
9] Wilhelms, D, E,, 1980, Strat igraphgy of part of the lunar
near side, U, S. Geol. Survey Prof. Paper 1046-A, 71p,
242
GOLDSTONE RADAR OBSERVATIONS OF MARS: THE 1986 OPPOSITION
Ts W^ Thompsonj Jet Propulsion Laboratory, Pasadena, CA 91109
Radar echoes from the planet Mars were obtained on 27 S-band (wavelength-12«5
cm) and 2 X-band (wavelength = 3,5 cm) tracks using the Goldstone Solar System
Radar. These observations took advantage of the favorable 1986 opposition
since the earth-Mars distance was 0,40 All at opposition (the smallest earth-
Mars distance since the 1971 and 1973 oppositions) and radar echo strength is
proportional to inverse-fourth -power to the distance to the target. Another
equally favorable opposition occurs in 1988; these favorable geometries do not
reoccur until the next century.
The coverages of the 1986 Goldstone radar observations are summarized in
Table 1 and Figure 1; which show the daily start and end point of each
observation. The observations were conducted via the cw-spectra techniques
described by Harmon et al, (1982 and 1985), A continuous tone was transmitted
at Mars and the radar echo was sampled to obtain a Doppler-spread spectrum.
Each received event was separated into polarized (opposite sense circular, OC)
and depolarized (same sense circular, SC) periods. Also, a minute or two of
noise was recorded in each transmit -receive cycle. The total echo time was
the round-trip travel -time which was varied from about seven minutes near
opposition to over twelve minutes for last runs in October, Thus, about one-
third of the total track was devoted to actual echo recording.
The coverage on Mars as shown in Table 1 started at 8° S, travelled toward
the equator to 3° S during August, arid then migrated south to 14° S for the
last run in October, These are new areas for earth-based radars. The data
analysis is just getting underway. However, our volatile real-time spectra
displays often showed features similar to those observed by Harmon et al,
(1982 and 1985),
There was one successful ranging run on 17 October 1986 (the last track). The
ground track for this run was similar to the cw observation of 15 October
1986, the Southern-most track in Figure 1, This ranging run had a resolution
of 2 microseconds and should yield surface heights accurate to 300 meters.
References :
J, K, Harmon, D, B Campbell, and S, J, Ostro (1982), Dual -Polarization Radar
Observations of Mars: Tharsis and Environs, Icarus , 52 , 171-187,
J, K, Harmon and S, J, Ostro (1985), Mars: Dual -Polarization Radar
Observations with Extended Coverage, Icarus, 62, 110-128,
243
^
+ DOY
DAY- ID
lUTC
LUTC
+FREQ+
-1-
+
+
LDN-LAT
ss
95.
131.
+ 169
860618
833
1102
S-BAND
9- 7S
+ 170
860619
1117
1255
S-BAND
4-
LON-LAT
=
126.
150.
9- OS
+ 174
860623
633
955
S-BAND
+
LON-LAT
=
21.
70.
8-30S
+ 175
860624
619
1225
S-BAND
+
LON-LAT
=
8.
98.
8-22S
+ 177
860626
611
1053
S-BAND
+
LON-LAT
=
349.
57.
8- 5S
+ 186
860705
946
1148
S-BAND
+
LON-LAT
=
321.
351.
6-40S
+ 187
860706
637
1121
S-BAND
+
LON-LAT
=
266.
335.
6-31S
+ 195
860714
452
801
S-BAND
+
LON-LAT
=
169.
215.
5- 168
+ 196
860715
530
930
S-BAND
+
LON-LAT
=
170.
228.
5- 7S
+ 198
860717
430
950
S-BAND
+
LON-LAT
=
137.
215.
4-50S
+ 201
860720
429
959
S-BAND
+
LON-LAT
=s
111.
191.
4-26S
+ 202
860721
434
824
S-BAND
+
LON-LAT
=
103.
159.
4-18S
+ 205
860724
410
850
S-BAND
+
LON-LAT
=
70.
139.
3-5BS
+ 209
860728
346
807
S-BAND
+
LON-LAT
=:
29.
92.
3-36S
+ 216
860804
319
752
S-BAND
+
LON-LAT
=
320.
26.
3-12S
+ 219
860807
302
826
S-BAND
+
LON-LAT
=
288.
7.
3- 8S
+ 221
860809
304
727
S-BAND
+
LON-LAT
=
271.
335.
3- 88
+ 228
860816
223
626
S-BAND
+
LON-LAT
ss
197.
256.
3-2 IS
+ 237
860825
445
653
S-BAND
+
LON-LAT
=
149.
180.
4- 6S
+ 241
860829
139
704
S-BAND
+
LON-LAT
=
66.
145.
4-35S
+ 244
860901
303
558
S-BAND
+
LON-LAT
=
59.
101.
5- OS
+ 246
860903
249
634
S-BAND
+
LON-LAT
=
37.
92.
5-19S
+ 250
860907
239
630
S-BAND
+
LON-LAT
=
357.
53.
5-5SS
+ 255
860912
430
627
X-BAND
+
LON-LAT
=
337.
5.
6-54S
+ 259
860916
211
643
S-BAND
+
LON-LAT
=
265.
331.
7-4 IS
+ 260
860917
225
626
X-BAND
+
LON-LAT
=
259.
318.
7-53S
+ 276
861003
420
523
S-BAND
+
LDN-LAT
=
135.
150.
11-2BS
+ 281
861008
335
600
S-BAND
+
LON-LAT
=:
76.
111.
12-408
+ 285
861012
2335
546
S-BAND
+
LON-LAT
329.
59.
13-508
+ DOY
DAY- ID
lUTC
LUTC
•i-FREQ+
Table
1: Goldstone Radar Obse
rvations
Of
Mars: 1986
50s-
LAT.
lOOS
150s
-I50S
180°
00/360°
270° 180°
LONGITUDE
27S-BAN0(X"13cm)RUNS
— 2 X-BANO (X ' 3.5 cm) RUNS
Figure 1: Latitude-Longitude Coverages of 1986 Goldstone
Radar Observations of Mars
244
A DIFFUSE RADAR SCATTERING MODEL FROM MARTIAN SURFACE ROCKS
W. M. Calviris B, M. Jakoskys Laboratory for Atmospheric and Space
PhysicSs University of ColoradOj and P. R. Christensenj Department of Geology,
Arizona State University
Remote sensing of Mars has been done with a variety of instrumentation at
a variety of wavelengths, Jakosky and Christensen (1986) have shown that many
of these data sets can be reconciled with a surface model of bonded fines (or
duricrust) which varies widely across the surface and a surface rock distribu-
tion which varies less so. Recently, a surface rock distribution map from -60
to +60° latitude has been generated by Christensen (1986). Our objective is
to model the diffuse component of radar reflection based on this surface
distribution of rocks. The diffuse, rather than specular, scattering is
modeled here because the diffuse component arises due to scattering from rocks
with sizes on the order of the wavelength of the radar beam. Scattering for
radio waves of 12.5 cm wavelength is then indicative of the meter-scale and
smaller structure of the surface. The specular term is indicative of large-
scale surface undulations and should not be causally related to other surface
physical properties. A model of the diffuse component could help us compare
various radar and infrared data sets and further constrain the nature of the
martian surface.
Based on the images of the Viking Lander sites and radar measurements,
diffuse scatterers do not appear to dominate the Martian surface. The
scattering particles are irregularly shaped and sized and may reside on top of
or within a dielectric discontinuity; this precludes a ready analytical solu-
tion to the scattering problem. Therefore, a simplified model of diffuse
scattering is undertaken.
Although it has been shown that multiple scattering by subsurface rocks
may make a significant contribution to the returned diffuse component in radar
scattering (Pollack and Whitehill, 1972), our simplified model assumes that
only the rocks on the surface will contribute. It is assumed that the rocks
are non-absorbing, so that all power extracted from the beam is scattered
(i.e., a single scattering albedo of one). Also, it is assumed that the power
is scattered isotropically, and the scattering efficiency, Q, is taken to be
one. This latter assumption is consistent with the Mie-scattering calcula-
tions of Hansen and Travis (1974) for particles of size parameter 1 to 6. The
returned power is normalized to that returned from a smooth planet, so that
common factors (e.g., incident power) divide out. The total power returned is
then proportional to the projected fractional rock coverage integrated over
the visible disk. Integration in one dimension, along lines of constant
doppler shift, can be performed to obtain the cross section as a function of
doppler shift, in a similar format to actual radar measurements.
There are two principle ways to express the rock distribution of a spher-
ical surface projected onto a disk. Further from the center of the disk the
surface rock distribution is viewed at an increasingly oblique angle. If the
rocks are sitting on the surface, the projected fractional surface coverage is
much higher at the limbs than it is normal to the surface, at disk center. We
take the apparent surface distribution to be given by f=l-exp(-T/cos i), where
T is a parameter used to fit the value of the rock abundance when viewed
normal to the surface and i is the angle between the surface normal and the
245
return beam» The angle i varies from to 90° over the face of the planetary
disk, so rock abundances vary from a nominal value (e.g.^ 10%) at the subradar
point (disk center) to 100% at the limbs of the planet. This is called^
henceforth, the 'exponential model'. Alternativelya the surface distribution
can be modeled as flat rocks imbedded in the surface^ the 'cobblestone'
model. Here^ the surface fraction covered by rocks does not depend on inci-
dence angle, and a planet of uniform coverage is represented by that uniform
value (e.g. a 10%) everywhere.
Both the cobblestone and exponential models were applied to a planet of
uniform fractional rock coverage with values ranging from 5 to 20%. This
yielded cross section versus Doppler shift curves which were reasonable in
shape, and total cross sections for the planet between 0.010 and 0.080,
depending on the rock abundance and model type. These values appeared reason-
able in light of the measured diffuse cross sections between 0.049 and 0.092
(Harmon and Ostro, 1985). Finally, we applied the exponential and cobblestone
models to the map of rock coverage (Christensen, 1986), and compared the
results to the published diffuse radar scattering curves (Harmon and
Ostro, 1985| Harmon, et al., 1982, hereafter, HO and HCO, respectively).
We found that although neither model fit the measured data, the models
gave values that were reasonable. The broad shape of the cobblestone model
was In reasonable agreement with the data. The magnitudes of the cross sec-
tion curves as well as the total cross section as a function of longitude were
lower than the values given by HO and HCO by a factor of 2 or 3; but given the
general assumptions of the model we did not expect to do better. The total
cross section as a function of longitude is also well correlated with the 2.5-
cm radio emission curve, which is to be expected because both the surface rock
map and the radio emission are correlated to thermal inertia (Jakosky and
Christensen, 1986).
The following aspects were in poor agreement with the data. The shape of
the exponential model was in serious disagreement with the data due to a large
degree of limb enhancement. Also, both models have a convex shape for large
doppler shifts whereas HO and HCO results are concave in this region. This
disagreement could be due to the effects of diffraction or multiple scattering
at highly oblique angles of incidence. Another disagreement was the lack of
duplication of small-scale features identified by HO and HCO, The locations of
these features, as determined by HCO, did not correspond to any obvious
features in the surface rock map. Also, there was no run-to-run correspondence
between the magnitudes of the modeled and actual cross section curves. This
may indicate the sensitivity of the actual measurements to scattering elements
in the subsurface or in the polar region, for which we have no data. Another
problem is the lack of uniqueness in the surface map. Twice as many scattering
elements of one-half the size would produce the same thermal contrasts, but
would have very different radar scattering properties. This implies the sur-
face map derived from thermal contrasts is dependent on the assumed size of
the scattering elements.
In an effort to bring the model into better agreement with the actual
measurements we are currently examining two possibilities. We plan to vary
the surface distribution of rocks in the polar regions to see if some of the
features reported by HCO could be accounted for. Also, we plan to vary the
surface distribution in the latitude and longitude bands which correspond to
246
the HCO features to see if a distribution which is consistent with the observ-
ed radar scattering as well as the thermal contrasts can be obtained,
REFERENCES
Christensens P.R.^ The Spatial Distribution of Rocks on Mars, Icarus , in
press s 1986.
Hansen, J,E, and L,D. Travis, Light Scattering in Planetary Atmospheres, Space
Sci. Rev, 16 , 527-610, 1974,
Harmon, J,K., D.B, Campbell, and S.J. Ostro, Dual-Polarization Radar Observa-
tions of Mars; Tharsis and Environs, Icarus 52 , 171-187, 1982.
Harmon, J.K. and S.J, Ostro, Mars: Dual-Polarization Radar Observations with
Extended Coverage, Icarus 62 , 110-128, 1985,
Jakosky, B.M. and P.R, Christensen, Global Duricrust on Mars: Analysis of
Remote-Sensing Data, J. Geophys. Res. 91 , 3547-3559, 1986,
Pollack, J,B, and L, Whitehill, A Multiple Scattering Model of the Diffuse
Component of Lunar Radar Echoes, J. Geophys. Res. 77 , 4289-4303, 1972.
247
MARS: SEASONALLY VARIABLE RADAR REFLECTIVITY
L. E. Roth, R. S. Saunders, and T. W. Thompson
Jet Propulsion Laboratory, Pasadena, CA 91109
Since reflectivity is a quantity characteristic of a given target at
a particular viewing geometry, the same (temporally unchanging) target
examined by radar at different occasions should have the same
reflectivity. Zisk and Mouginis-Mark (1980) noted that the average
reflectivities in the Goldstone Mars Data (Downs et al., 1975) increased
as the planet's S hemisphere passed from the late spring into early
summer. We have examined the same data set and confirmed the presence in
the data, of the phenomenon of the apparent seasonal variability of
radar reflectivity (Roth et al. , 1984; 1985). Objections were raised
against our reports. These objections fell into three categories:
(1) Reflectivity variations may be present in the Goldstone Mars data.
Their presence must be the result of an instrumental/calibration
error.
(2) Reflectivity variations may be present in the Goldstone Mars data.
Since there is a two-year interval between the two experiments,
the variations must be the result of differences in the data
reduction procedures applied first to the 1971 data and then to
the 1973 data.
(3) Reflectivity variations are not present in the Goldstone Mars
data. The variations were - Introduced into the analysis through
comparing reflectivities obtained during two separate experiments .
In other words, what appears to be a seasonally variable
reflectivity is, in fact, the result of a joint analysis of two
incompatible subsets of the combined data set.
Our work in FY' 86 was mostly aimed at answering the listed
objections. We have completed the effort aimed at validating the
Goldstone (1971, 1973) Mars data set. We have reviewed the procedures
followed during both the 1971 and 1973 Goldstone Mars experiments and
examined the available records . We present here a summary of the
principal findings pertaining to Objections (1) and (3).
System calibrations were a regular feature of each observing run.
Included in the calibrations were: (1) Measurements of the system
temperature. (2) Measurements of the antenna gain variations vs.
elevations. (3) Measurements of the antenna gain variations due to
structural modifications. (4) Measurements of the transmitter power.
Early in each opposition transmitter calibrations were performed at the
start and at the end of each run. Later, when no drifts were observed,
the transmitter calibrations were discontinued. (5) Measurements of
the additive noise of the microwave links (when links used). (6)
Monitoring of the antenna pointing accuracy. Tracks of calibration
radio sources were regularly scheduled and the results were folded into
the operational procedures. We estimate the resulting total
reflectivity calibration error to be less than 5% of the respective
absolute values .
248
After having applied all the known corrections to the radar system
sensitivity, global reflectivity averages were computed for each
opposition. The results are:
<R>(1971) = 0.0564
<R>(1973) = 0.0625
<R>(A11 data) = 0.0594
It is seen that the average reflectivity in the 1973 data (240
deg<Ls<325 deg) is higher than the average reflectivity in the 1971 data
(200 deg<Ls<275 deg), in agreement with Zisk and Mouginis-Mark (1980)
and Roth et al., (1984; 1985). Note that the ratio
<R>(1973)-<R>(1971)
<R>(A11 data)
is equal to 11%, a value twice that of the estimated calibration error.
This discrepancy could be interpreted in the following manner: (1) The
difference in the reflectivity averages in the 1971 and 1973 data is
being caused by some unknown and unaccounted for instrumental error and
calibration drift. This error is of the approximately same magnitude as
all the known uncertainties. (2) The difference in the reflectivity
averages in the 1971 and 1973 data is being caused by differences in
coverage. The overlap is sparse and thus the difference in the mean
1971 and 1973 reflectivities could be caused by differences in coverage.
(3) The difference in the reflectivity averages in the 1971 and 1973
data is being caused by changes in the target characteristics. Those
changes may be caused by two agents: dust precipitation/removal or
thawing of subsurface ice. Modeling exercises indicate that a shifting
dust cover is not likely to be a major contributor to the observed
reflectivity variations (Zisk and Mouginis-Mark, 1981; Roth et al. ,
1986a) . The liquid-water hypothesis has not been supported by a
credible model of the thermal regime in the upper 1 m of the Martian
surface. Thus all three interpretations are about equally likely or
unlikely, depending on the point of view. The liquid-water hypothesis
could, in principle at least, account for the pattern of seasonal
variability, whereas the other interpretations could not.
To address Objection (3), we investigated the statistical
relationship between reflectivities and the areocentric longitude, Ls ,
separately for the 1971 and 1973 data (Roth et al., 1986b). The
computations were carried out separately for the 1971 data, the 1973
data for the cases when the 1973 scan was taken at a higher solar
longitude as the 1971 scan, and for all data combined. Conclusions:
(1) Correlation coefficients between the mean reflectivity ratios and
the lengths of the temporal separation of overlapping scans are positive
for the 1971 Goldstone Mars data. This means that the average
reflectivities in the 1971 data tend to increase as the S hemisphere
passes from the vernal equinox to the summer soltice. (2) Correlation
coefficients between the mean reflectivity ratios and the lengths of the
temporal separation of overlapping scans is largely negative for the
1973 Goldstone data. The mean reflectivities in the 1973 data appear to
undergo a mild decrease as the S hemisphere enters late summer. At
249
Crit=0.3 (for the definition of the quantity Grit see Roth et al. ,
1986b) there is an exception to this general trend. This exception is
significant in that it shows that the data set is statistically
inhomogeneous . Random removal of a few elements from the sample affects
the sample statistics. Thus any conclusions based on purely statistical
arguments have to be received with caution. Statistical inhomogeneity
of the data deserves further investigation. (3) The apparent seasonal
pattern in the behavior of the mean reflectivities is not the result of
the joint analysis of the 1971 and 1973 data. This is our most
important finding. The seasonal reflectivity variations may or may not
be real. However, they are certainly a characteristic property of each
individual subset, rather than of the combined (1971, 1973) Goldstone
set. (4) Reflectivity variations in the 1971 data are consistent with
the hypothetical presence of subsurface moisture passing through a
seasonal freeze-thaw cycle. If the correlation coefficient for Crit=0.3
is ignored, reflectivity variations in the 1973 data are also consistent
with the liquid water hypothesis, provided we accept a naive, intuitive
notion that subsurface moisture in the subequatorial areas freezes after
the S hemisphere passes the summer stoltice. (5) Consistency is not
equivalent to a proof. Reflectivity variations could only be considered
a proof of the existence of the subsurface moisture in the equatorial
areas of Mars if the characteristic pattern of seasonal behavior were to
be confirmed by further, preferably multifrequency, radar observations.
Downs, G. S., R. M. Goldstein, R. R. Green, G. A. Morris, and P. E.
Reichley (1973). Martian topography and surface properties as
seen by radar: The 1971 opposition. Icarus 18 , 18-21.
Roth, L. E., R. Saunders, and G. Schubert (1984). Radar and the
detection of liquid water on Mars (abstract) . In Abstracts for
Water on Mars Workshop, 63-65. Lunar and Planetary Institute,
Houston.
Roth, L. E., R. S. Saunders, and G.
variable radar reflectivity.
Schubert (1985).
Lunar Planet. Sci.
Mars : Seasonally
XVI, 712-713.
Roth, L. E., R. S. Saunders, and T. W. Thompson (1986a). Radar
reflectivity of a variable dust cover. In Abstracts for Dust on
Mars Workshop, 59-61. Lunar and Planetary Institute, Houston.
Roth, L. E., R. S. Saunders, G. S. Downs, and G. Schubert (1986b).
Mars: Seasonally variable radar reflectivity II. Lunar Planet.
Sci. XVII, 730-731.
Zisk, S. H. and P. J. Mouginis-Mark (1980). Anomalous region on Mars:
Implications for near-surface water. Nature 288, 735-738.
Zlsk, S. H. and P. J. Mouginis-Mark (1981). Alternate models for the
Soils Lucas radar anomaly on Mars (abstract). In 3rd Int'l
Colloq. Mars, 294-296. Lunar and Planetary Institute, Houston.
250
MEASUREMENTS OF THE DiELECTRIC CONSTANTS FOR PLANETARY
VOLATILES.
Vincent G. Anicich and Wesley T. Huntress, Jr.
Jet Propulsion Laboratory, Pasadena, CA 91109
The model of Titan at present has the surface temperature, pressure,
and composition such that there is a possibility of a binary
ethane— methane ocean (ref. 1). Proposed experiments for future Titan
flybys Include microwave mappers. Very little has been measured of the
dielectric properties of the small hydrocarbons at these radar frequencies.
Nor, Is there much known about the loss tangent (Imaginary component of
the dielectric constant) which relates to the penetration distance of the
microwaves Into the liquids.
We have set up a laboratory experiment, utilizing a slotted line, to
measure the dielectric properties of the hydrocarbons, methane to heptane,
from room temperature to —180 C. Temperatures below 25 C are
maintained using various ice baths. A large dry box Is used to eliminate the
condensation of water vapor at these lower temperatures.
The literature reveals very little data on these systems. A substantial
study was made into the real part of the dielectric constants of these
hydrocarbons using a 1 kHz frequency (refs. 2 — 6). Besides being so far away
from 1.2 GHz, the proposed frequency of the mappers, the measurements
were made at both reduced and elevated pressures. The change in pressure
has a marked effect on the magnitude of the dielectric constants and
therefore these earlier results must be evaluated.
Figure 1 is a graphical summary of our results thus far. We have thus
determined that our experimental measurements of the real part of the
dielectric constants are accurate to ±0.006 and the imaginary part of the
dielectric constants, the loss tangent, of the liquids studied is 1.001. In
order to verify this low of a loss tangent we studied the real part of the
dielectric constant of hexane at 25 C as a function of the frequency range
of the slotted line system that we have. Figure 2 show these results. The
dielectric constant of hexane at room temperature, between 500 MHz and
3 GHz, is constant within our experimental error. The real parts of the
dielectric constants measured here are consistent with the those that can
be compared to the previous literature after pressure extrapolations.
REFERENCES
1. Lunine. J. I.. Stevenson, D. J., and Yung. Y.L.. 1983, Ethane Ocean on
Titan, Science, 222, p. 1229-1230.
2. G. C. Straty and R. 0. Goodwin, 1973, Cryogenics , Dec, p. 712.
3. W. P. Pan, M. H. Mady, and R. C. Miller, 1975, AlChE Journal, 21, p.
283 .
4. R. T. Thompson, Jr. and R. C. Miller, 1980, Adv. Cryoq. Enq ., 25, p. 698.
5- W. G. S. Scaife and C. G. R. Lyons, 1980, Proc. R. Soc. Lond., A370, p.
193.
5. M. G. Gaikwad. R. Chandkrasekar, S. K. David, and V. G. Alwani, 1980,
Phys. Lett. , 80 A , p. 201.
251
Figyre 1.
Dielectric Constants of Hydrocarbons, a 1.2 GHz
c
(0
+J
M
C
o
o
o
•H
o
0)
H
Q)
•H
a
2.10
2.05 h
2.00
1.95
1.90
1.85
1.80
-120
JL
I
-80 -40
Temperature, C
40
Heptane
o
Hexane
Pentane
Butane
Propane
Floure 2.
Dielectric constant of Hexane, a Z5 C
2.00
1.95
•p
1.90
c
(0
■p
c
1.85
o
o
u
1.80
•H
».<
+J
1.75
H
(U
.H
a
1.70
.0
1.0 1.5 2.0
Frequency, GHz
2.5
3.0
252
APPLICATION OF NUMERICAL METHODS TO PLANETARY RADIOWAVE SCATTERING
Richard A« Simpson and G. Leonard Tyler
Center for Radar Astronomy
Stanford University » Stanford^ CA 94305
Continued advances in computer technology make numerical solution of
certain scattering problems feasible. Our work includes investigation of
existing techniques to determine those which might be applicable to planetary
surface studies, with the goal of improving the interpretation of radar data
from VenuSj Mars, the moon, and icy satellites.
Numerical scattering models currently require that the surface be
approximated by flat, perfectly conducting facets -- as shown in the figure^
Each facet is described in some standard way, such as by its unit surface
normal and its area. Facet dimensions are limited to fractions (typically
one-tenth) of the probing wavelength. The total field is found by
simultaneously solving the boundary condition problem on all the facets which
make up the approximation to the surface. These equations relate the incident
fields the reradiated (scattered) field, and the currents.
Because the complexity of the computation increases rapidly with the
number of facets, only relatively simple surfaces have yielded numerical
solutions. Somewhat larger structures may be studied if symmetries -- such as
about a rotation axis -- are identified, but these are not helpful in the case
of randomly rough surfaces. The simple geometrical shapes sometimes have
analytical scattering solutions. Comparison of the numerical and analytical
results provides Insights into the strengths and limitations of both methods.
After experimenting with some smaller numerical codes, we are now
examining the Numerical Electrogmatics Code (NEC) developed at the Lawrence
Livermore Laboratory, This is one of the most powerful, publicly available
codes for scattering analysis in three dimensions. Though not developed for
random rough surfaces, it contains elements which may be generalized and which
could be valuable in the study of scattering by planetary surfaces.
Figure - Gently undulating surface is approximated by__facets d^efined by
unit surface normals n. Orthogonal current components J, and Jp link
incident and scattered fields through boundary conditions at the surface.
253
CHAPTER 8
AEOLIAN PROCESSES
BRIGHT SAND/DARK DUSTs THE IDENTIFICATION OF ACTIVE SAND
SURFACES ON THE EARTH AND MARS
H.G.Blount IP, R'Greeley^, P.R.Christensen^ and R. Arvidson^, ^Dept of
Geology, Arizona State University, Tempe, Arizona, 85287, ^Earth Sciences/ Geology Dept.,
Washington Univiversity, St. Louis, Mo. 63130
Aeolian features are common on Mars. Although wind streaks are known to be active
features, controversey has arisen regarding the activity of sand sheets and dunes, particularly in
the north polar erg. In the absense of in situ measurements, remote sensing techniques must be
relied upon to address the problem. Field studies and analysis of LANDSAT Thematic Mapper
data in the Gran Desierto, Mexico (Blount et al., 1986) may shed light on a technique to
distinguish active from inactive (relict) sand surfaces. Active sand bodies are here defined as
those with saltation surfaces on which wind-induced ripple marks and/or shpfaces are observable
in field outcrop. Inactive sands are those surfaces which do not show evidence of wind
movement; sand drapes and fluvial sands being the best examples. Active sand bodies in the
study area are consistently brighter (by an average of 20%) at visual and near-infrared
wavelengths (0.45-2.35|i.m) and darker at thermal infrared wavelengths (lG.3-12.5M.m) than
compositionally similar inactive sands. The Gran Desierto study area covers more than 5,500
km2 and includes major transverse, longitudinal and star dunes.The area is the largest sand sea in
North America and has been described in detail by Lancaster et al. (in press). Sands are primarily
of quartz composition (80-99%) with varying amounts of amphibole, biotite, carbonate
fragments and volcanic detritus as assessory grains.
The reasons for the albedo difference between active and inactive sands are partly explained
by a textural analysis of the two terrains. Active surfaces are composed of relatively higher
percentages of saltation-sized grains; their higher reflectivity is in agreement with Gerbermann
(1979) who found that reflectivity increased directly with weight % sand in sand/silt/clay
mixtures. Material in the coarse size fraction (>250p.) is dominated by fragments of amphibole
and opaque volcanic hthics (basalt). This size-fraction is generally absent from samples of active
sand but makes up as much as 5-10 weight % of many inactive samples. Saltation-sized particles
(62.5-250)1) are primarily quartz with minor amphibole and biotite. Active sands are highly
unimodal (90-99 wt.%) in this size range. The silt-sized and smaller fraction (<62.5|i) also
contains abundant quartz but with relatively more amphibole and aggregated opaque lithics
present. Active sands are depleted with respect to inactive sands in boA the smallest and largest
size fractions present. Inactive sands contain small but conspicuous amounts of coarse sand and
silt; the sizes which are dominated by darker materials. It has long been known that minute
inclusions of dark material can drastically reduce the reflectance of a mixture (Singer, 1981;
Clark, 1983) leading to the conclusion that the apparent darkness of inactive sands is caused by
the inclusion of these darker size fractions in small quantities. In the northeastern portion of the
Gran Desierto however, darker inactive sands were observed which also lacked a coarse-size
fraction; all of the darkening being attributeable to the higher weight percent of darker silt-sized
grains. In the TM images, low albedo (at VNIR wavelengths) corresponds to regions of higher
silt/clay content. After seiving at one phi intervals both active and inactive samples show a
decrease in visual reflectance with particle size. The apparent darkening of these sands with
decreasing grain size is apparentiy contradictory to the characteristic silicate behavior of
increasing reflectance with decreasing grain size (Sahsbury and Hunt, 1968). While the "normal"
situation holds true for the individual components of Gran Desierto sands (primarily quartz and
amphibole), it does not describe the mixture of the two end-members. As darker litMc fragments
break down to smaller grain sizes, their effective surface area increases in relation to the quartz
component. Decreasing grain size then results in a lowering of the overall albedo for the mixture.
These results are compatible with laboratory spectral work by Clark (1983), who showed that the
overall reflectance in a Ught/dark mixture (montmorillonite + charcoal) decreases with the particle
size of opaques, though the effect was attributed to opaque coatings on otherwise bright grains
257
rather than the increased surface area of opaques present in smaller particle sizes. Opaque coating
have not been observed in the Gran Desierto samples however inactive sand samples always
contain quartz/Uthic aggregates ia their smallest size fraction (<62.5|X). Saltation-size particles of
opaques are depleted in both the active and inactive samples. AeoUan sorting may be responsible
for this phenomenea, although the presence of Uthic opaques in the silt/clay fraction indicates that
any saltation-sized fragments are rapidly abraded into smaller-sized grains.
The mixing model of Johnson et al. (1983) has been investigated for tracing the provenance
of active and inactive sands based on albedo and spectral variations. If work in other field areas
(Kelso Dunes, Ca. and Mohawk Dunes, Az.) confirms that each size fraction indeed possesses a
unique reflectivity (assuming constant bulk composition) then a self-mixing model may develop
in which each size fraction of the mixture can be considered an independent end-member. It is
therefore not simply the weight percent of opaques, but also the weight percent and cross-
sectional area of a given size fraction present wWch will control the total reflectance of a sand/silt
mixture. Active sands are found to be brighter than inactive sands, not only because they contain
fewer opaques but because the opaques they contain are larger in size and are contributing less to
the total sand body radiance.
The Mars Observer (scheduled for launch in 1990) will image the surface at wavelengths
from 0.4 to 5.2 ^im with the Visual Near-Mrared Mapping Spectrometer (VIMS) and from 6.25
to 50 |im with the Thermal Emission Spectrometer (TES). Portions of these wavelength regions
correspond to the Thematic Mapper data which was used to detect albedo differences between
active and inactive surfaces. The identification of active sands on Earth, with a priori knowledge
of bulk composition and grain size distribution, may allow the remote mapping of active sand
surfaces on Mars. In conjunction with thermal infrared remote sensing for composition, it may
also provide a method for the remote determination of grain size distributiuons within sand/silt
mixtures.
REFERENCES:
Blount, H.G.II, Greeley, R. and Christensen, P,R., 1986, Geol. Soc. Amer.Abstr with Prog.,
V. 18, n. 6
Clark, R.N., 1983, J. Geophys. Res., v. 88, p.10635-10644
Gerbermann, A.H., 1979, Photogrammetric Engr. and Remote Sensing , v. 45, p. 1145-1151
Johnson, P.E., Smith, M.O., Taylor-George, S. and Adams, J.B., 1983, J. Geophys.Res.,
V.88, p. 3557-3561
Lancaster, N,, Greeley, R. and Christensen, P.R., in press, Earth Surface Processes and
Landforms
Salisbury, J.W. and Hunt, G.R., 1968, Science, v.l61,p.365-366
Singer, R.B., 1981, J. Geophys. Res., v.86, p. 7967-7982
258
REGIONAL SOURCES AND SINKS OF DUST ON MARSi VIKING
OBSERVATIONS OF CERBERUS, SOLIS PLANUM, AND SYRTIS MAJOR
Steven W. Lee, Department of Geology, Arizona State University, Tempe, AZ 85287
Seasonal variability of classical martian albedo features has long been noted by terrestrial
observers [1,2]. Spacecraft observations of such features have shown them to be related to
aeolian transport of bright dust into and out of regions, primarily in association with major dust
storms [cf. 3,4]. Investigation into the amount and direction of dust transport related to variable
features can reveal regions which, at present, act as either sources (net erosion of dust from an
area) or sinks (net deposition) [5,6].
A study of seasonal variations of albedo features in the Cerberus, Solis Planum, and
Syrtis Major regions has been based on Viking Orbiter data obtained over more than one complete
martian year. Contour maps of Lambert albedo and single-point thermal inertia have been
constructed from the Infrared Thermal Mapper (IRTM) experiment data, and Orbiter images have
been used to determine the pattern and variability of regional winds ( inferred from wind streak
orientations). Coupled with ground-based radar data, these data sets ailow the regional sediment
transport direction, surface properties (texture, morphology, and roughness), and 5ie implications
of the observed seasonal and longer term dust redistribution, to be investigated. The results of
this study are outlined below.
Solis Lacus, the most prominent dark albedo feature in Solis Planum, extends over
approximately 20° of longitude and 10° of latitude (centered at 25°S, 85°W), and contains and is
surrounded by a conspicuous pattern of bright and dark wind streaks. The albedo feature is
highly variable in extent and contrast with its surroundings, generally being most distinct during
southern spring and summer (minimum Lambert albedo ~0.13) and less distinct during southern
fall and winter (minimum albedo ~0.16). The regional thermal inertia values (~8-10 x 10'^
cal/cm2/seci^2/oj^) ^g indicative of a surface covered by particles larger than ~100 iim [7]. A
seasonal dust- transport cycle has been proposed to explain these observations: 1) During late
southern spring and summer, bright dust is eroded from the surface (possibly ejected by a
saltation triggering mechanism) and transported from the region by local dust storms (several were
detected by Viking, and have been commonly observed from Earth). Removal of dust over a
wide area results in the dark, distinct, Solis Lacus feature. 2) Following cessation of dust-storm
activity, sedimentation from the atmospheric dust load occurs over the entire region, decreasing
the contrast of the albedo feature with its surroundings. 3) The cycle may be renewed by dust-
storm activity the following year. The retention of some albedo features throughout the year, plus
the constancy of the regional thermal inertia, requires that the albedo features do not involve
erosion or deposition of substantial deposits; cycling of, at most, a few tens of ^.m of dust is
indicated. Differences in time of occurrence, severity, and longevity of dust-storm activity may
lead to the observed year-to-year changes in Solis Lacus.
Dramatic seasonal variability is also characteristic of Syrtis Major (~5°S - 25°N, 275° -
300°W). The feature hes on the low-albedo slopes of a volcanic shield [8] (generally darker than
~0.2 in albedo), the darkest area (albedo ~0.1) being closely associated with a mass of dunes
located near the crest of the shield. Thermal inertias of ~8 x 10'^ cal/cm^/sec^^^/oK^ plus the
observed dunes, argue strongly for a sandy surface. Syrtis Major increases in altedo immediately
following global dust storms, then darkens steadily through the balance of the year until reaching
its pre-storm albedo (also confirmed in [9]). The observed trend of bright and dark streaks is in
response to winds generally directed upslope and to the west. The dust-transport cycle consistent
witii these observations is: 1) Enhanced deposition from global dust storms increases the regional
albedo. 2) The relatively mobile surface coupled with effective regional winds (possibly
reinforced by the global circulation) results in ejection of dust from the surface and net transport
to the west during the remainder of the year, yielding a decreased regional albedo. 3) Beginning
of another global dust-storm cycle begins the process again. Such a transport cycle provides a
259
mechanism for significantly enhanced deposition in the neighboring low thermal inertia region,
Arabia (as suggested in [10]).
Cerberus is located to the south of the Elysium volcanoes on the gentle slopes (< 0.5°) of
Elysium Planitia, extending over about 10° of latitude and 20° of longitude. The feature is
generally darker than ~0.2 in albedo, with the eastern portion being darker (minimum albedo of
~0.14) than that to the west (minimum albedo -0.16). The albedo pattern is closely correlated with
the regional thermal inertias, with the darkest areas having inertias of ~10 x 10'^
cal/cm2/seci^2/oj^^ while the slightly brighter western region exhibits lower inertias of ~8. The
entire Cerberus feature is surrounded by brighter (albedos to 0.30), lower inertia (~4) material.
The pattern of wind streaks suggests relatively constant effective wind directions from the east,
indicating that the gentle topography is not sufficient to significantly alter or enhance the global
circulation [5]. Seasonal albedo variations closely mimic those observed in Syrtis Major,
although the post-dust-storm darkening is not as pronounced, suggesting that the aeolian
environment is less vigorous in Cerberus. The albedo and thermal inertia east-west asymmetry is
consistent, however, with coarser material residing in the east and being transported westward by
the global winds. Cerberus may therefore be acting as a source of sediment which is transported
by saltatioii rather than in suspension. .
These observations indicate that very different levels of aeolian activity give rise to the noted
seasonal variability of these regions: 1) Syrtis Major presents a picture of an extremely active
aeolian environment. Deposition occurs only during and immediately following major dust
storms, while during the rest of the year regional winds are apparentiy effective agents of dust
ejection and transport. Syrtis Major thus acts as a dust source region through most of the year,
possibly supplying the neighboring dust sink region of Arabia. 2) Cerberus may be thought of as
a less-vigorous version of Syrtis Major, with definite removal of dust following global storms. A
net westward flux of coarser material is evident, possibly encroaching onto surrounding Elysium
Planitia. 3) In Solis Lacus, local dust storms result in net removal of dust from the area, while
deposition occurs from the atmospheric dust load throughout the remainder of the year. Solis
Planum serves as a source region of dust only during the limited period (dust storm season) when
regional winds are capable of inducing dust ejection from the surface. Contributions to the global
dust load may occur only during major dust storms.
REFERENCES
[1] Slipher, E.C., 1962, A Photographic History of Mars, 1905-1961, Nortiiland Press,
Flagstaff, AZ.
[2] De Mottoni, G., 1975, Icarus 25, 296-332.
[3] Sagan , C. et al., 1973, /. Geophys. Res. 78, 4163-4196.
[4] Thomas, P. and J. Veverka, 1979, /. Geophys. Res. 84, 8131-8146.
[5] Lee, S.W. et al, 1982, /. Geophys. Res. 87, 10025-10041.
[6] Lee, S.W., 1986, submitted to Icarus.
[7] Kieffer, H.H., et al., 1977, /. Geophys. Res. 82, 4249-4295.
[8] Schaber, G.G., 1982, /. Geophys. Res. 87, 9852-9866.
[9] Christensen, P.R., 1986, Lunar Planet. Sci. Conf.XVII, 121-122.
[10] Christensen, P.R., 1982, /. Geophys. Res. 87, 9985-9998.
260
HIGH RESOLUTION THERMAL INFRARED MAPPING OF MARTIAN
CHANNELS
RA, Craddock, R. Greeley, and P.R. Christemsen, Department of Geology and
Center for Meteorite Studies, Arizona State Univesity, Tempe, Arizona 85287.
Morphologic studies [e.g., 1,2] have not yielded a unique interpretation for channel
formation [3-7] because of the ambiguity between geomorphologic features and their
similarities of one mode of formation with another. For example, teardrop shaped islands,
or streamlined hills, have been used in support for all of the channel formation hypotheses
and could easily be explained by any of them. Viking Infrared thermal mapper (IRTM)
high resolution (2 to 5-km) data have been compiled and compared to Viking Visual
imaging subsystem (VIS) data and available 1:5M geologic maps for several martian
channels including Dao, Harmakhis, Mangala, Shalbatana, and Simud Valles in an effort to
determine the surface characteristics and the processes active during and after the formation
of these channels.
Mangala Valles is an approximately 850-km long channel located in the Menmonia
quadrangle that cuts through the southern cratered plains from 18.5°S, 149.5°W to
Amazonis Planitia at 4.0°S, 150.0°W. Its complicated pattern has been suggested as being
the result of several channeling episodes [8]. Channel deposits have been mapped as being
approximately 10 to 20-km in width to over 100-km towards the distal portion of the
channel [9]. Shalbatana and Simud Valles debouch into Chryse Planitia in tiie Oxia Palus
quadrangle at approximately 10°N, 40°W. Shalbatana VaUis is a sinuous channel with one
large tributary and originates in chaotic terrain at 1.0°S, 45.5°W extending NNE for
approximately 650-km. Its mapped channel deposits are between 10 to 25 -Ian in width
[10]. Simud VaUis originates in Hydroates Chaos at 2.0°N, 36.0°W and may have deposits
extending from 20 to over 250-km in width [10] while reaching over 400-km in length.
Both Dao and Harmakhis Valles debouch into Hellas Planitia in the Hellas quadrangle.
Dao Vallis is a sinuous channel originating in chaotic terrain 33.0°S, 266.0°W and extends
for over 500-km to approximately 40.0°S, 275.0°W. Its mapped channel deposits extend
from approximately 5 to 50-km in width [11]. Harmakhis Vallis is also sinuous and
originates in chaotic terrain closely associated to ReuU Vallis at 39.0°S, 265.0°W.
Harmakhis Vallis extends for over 450-km and has channel deposits mapped as being 5 to
over 40-km wide [11].
This study has involved the use of the highest resolution IRTM data tracks available
occurring at ideal conditions (range, to 8,(X)0-km; emission angle, 0° to 60°; time of day,
22 to 2 H; Lg, 0° to 2(X)°). Typically at least two high resolution data tracks occurring at
different intervals across the channels' distal extent were examined for each of the above
channels; this was needed in an effort to understand the distribution and possibly the
mechanism(s) for deposition of material within the channels. Channels such as Huo
Hsing, ReuU, Samara, and Tiu Valles, among others, have been excluded from this study
due to a lack of this optimum data.
Our results show a dominance of aeolian processes active in and around the channels.
These processes have left materials thick enough (lO's of cm) to mask any genuine channel
deposits. High resolution thermal inertia observations taken across the channels' distal
extent in various locations compare well with moderate resolution values determined for the
corresponding local surrounding regions [12]. A few of the channels (Dao, Harmakhis,
and Mangala Valles) showed an increase in thermal inertia with increasing distal extent as
observed at Kasei Vallis [13], but these increases were not unique to the area within the
channel itself and could be explained as a result of regional trends in grain-size distribution.
261
Moreover, none of the channels included in this study showed a higher thermal inertia than
their surroundings as observed at Ares [13], Ma'adim, and Al-qaMra Valles [14]. Results
from [13,14] indicate that certain martian channels act as topographic barriers trapping high
thermal inertia materials, but instead, our results indicate that very comparable martian
channels and their surrounding terrain are blanketed by deposits which are homogeneous in
their thermal inertia values. However, optimum IRTM data does not cover the entire
martian surface and because local deposits of high thermal inertia material may not be large
enough in areal extent or may be in an unfavorable location on the planet, a high resolution
data track may not always occur over these deposits. Therefore, aeolian processes may be
even more active than the IRTM data tracks can always show. Observations of Ma'adim
and Shalbatana Valles illustrate this point
Ma'adim and Shalbatana Valles are comparable in length (over 900-km versus over
650-km respectively), width (10 to 25-km for both ), slope (1/2° to 1° for both), and
elevation (from at the distal portion to over 4-km up-channel). They differ in terms of
latitude, geologic units, local wind streak patterns, and slightiy in the surrounding albedo.
Ma'adim Vallis is in the southern hemisphere beginning at 28.0°S, 183.0°W and extending
north to 16.0°S, and Shalbatana Vallis is in the northern hemisphere (see above). Ma'adim
Vallis is located in hilly and cratered material, which is interpreted as being heavily
impacted ancient terrain [15], and Shalbatana is in cratered plateau material, which is
interpreted as being heavily impacted lava flows [10]. Local wind streak patterns for
Ma'adim Vallis/ Aeolis quadrangle are towards the ESE; Shalbatana Vallis/Oxia Palus
quadrangle are towards the SSE. Ma'adim is located in a large low albedo region that
extends from the Aeolis quadrangle west into Hesperia Planum and further [16,17].
Corresponding wind streak patterns indicate that this low albedo material could move into
Ma'adim Vallis and surrounding local craters as suggested in [14,18].
Shalbatana Vallis is approximately 800-km south of a low albedo region situated in
northern Chryse and Acidalia Planitias [16, 17]. Based on observed wind streak patterns,
we suggest that this low albedo material could move south and remain trapped in
Shalbatana Vallis and surrounding large craters. The Oxia Palus quadrangle has dark
material centered at 8.0°N, 42.5°W in the junction between Shalbatana Vallis and its
tributary, as well as dark material in the south sides of large local craters [19]. However,
whether this material is similiar to that observed in Ma'adim and Al-qahira Valles is difficult
to determine because the highest resoultion IRTM data tracks do not cover these deposits.
The high thermal inertia area reported in nearby Ares Vallis [13] corresponds well to a dark
deposit centered a 8.5°N, 22.0°W [20]. This supports the possibility of unobserved high
thermal inertia deposits due to the position of the high resolution IRTM data tracks. Our
study has also shown that with optimum available data, the mechanism(s) for channel
formation will continue to remain ambiguous due to aeolian deposits masking the surface.
Future studies should concentrate on the other (i.e., less than optimum) high and additional
moderate (30 to 35-km) resolution IRTM data so that the remaining channels may be
investigated.
References
[1] Sharp, R.P., and M.C. Malin, 1975, Geol. Soc. Amer. Bull., 86, 593-609.
[2] Masursky, H., et al, 1977, Jour. Geophys. Res., 82, 4016-4038.
[3] Baker, V.R., and D.J. Milton, 1974, Icarus, 23, 27-41.
[4] Cutts, J.A., and K.R. Blasius, 1981, Jour. Geophys. Res., 86, 5075-5102.
[5] Lucchitta, B.K., et al, 1981, Nature, 290, 759-763.
[6] Nummedal, D., 1978, NASA TM 79729, 257-259.
[7] Schonfeld, E., 1976, (abst), Eos Trans. AGU, 57, 948.
[8] Masursky, H., et al., 1986, (abst). Lunar Planet. Sci. XVII, 520-521.
[9] Mutch, T.A. and E.G. Morris, 1979, U.S.G.S., Map 1-1137 (MC-16).
262
[10] WUhelms, D.E., 1976, U.S.G.S., Map 1-895 (MC-11).
[11] Potter, D.B., 1976, U.S.G.S, Map 1-941 (MC-28).
[12] Christensen, P.R., 1986, Jour. Geophys. Res., 91, 3533-3545.
[13] Christensen, P.R., and H.H. Kieffer, 1979, Jour. Geophys. Res., 84, 8233-8238.
[14] Zimbelman, J.R., 1986, (abst), Papers Presented to the Symposium on Mars:
Evolution of Its Climate and Atmosphere, 112-114.
[15] Scott, D.H., et al., 1978, U.S.G.S., Map I-llll (MC-23).
[16] Kieffer, H.H., et al., 1977, Jour. Geophys. Res., 82, 4249-4291.
[17] U.S.G.S., 1981, Map 1-1343 (MC-11 SW).
[18] Christensen, P.R., 1983, Icarus, 56, 496-518.
[19] U.S.G.S., 1981, Map 1-1343 (MC-11 SW).
[20] U.S.G.S., 1981, Map 1-1342 (MC-11 SE).
263
MARS: MORPHOLOGY OF SOUTHERN HEMISPHERE INTRACRATER
DUNEFIELDS.
N. Lancaster and R. Greeley, Department of Geology, Arizona State University,
Tempe, Arizona 85287.
Although most eolian dunes on Mars occur in the North Polar sand sea (Breed et al., 1979;
Tsoar et ^., 1979), there are important areas of dunes at mid to high latitudes in the southern
hemisphere. At least 75% of these dunes occur within craters (Tliomas, 1981; Peterfreund,
1985). In addition to the areas of dunes, many crater floors are mantled by material of
probable eolian origins (Thomas, 1981; Christensen, 1983). A large cluster of intracrater
dunefields lies in Noachis Terra between latitudes -40 to -48° and longitudes 330-340°, with
major dune accumulations in the craters Kaiser, Proctor and Rabe. Dunes in this region were
amongst the first to be recognised on Mars (Cutts and Smith 1973; Breed 1977), but have not
been studied in detail on ViMng Orbiter images.
The size of the dunefields ranges from 40 to 3600 km^ and varies directly with crater size.
Preliminary studies of Viking Oribiter images of intracrater dunefields reveals that they consist
of two varieties. The most common type is composed of massed straight to slightly wavy
crescentic dunes similar to those described by Breed (1977) and Breed et al. (1979). Dune
fields of this type occupy more than 20% of the area of the crater floor, with the dunefield
margins often marked by a large "dune wall" or rampart. Dune spacing ranges between 0.7
and 1.2 km. Slipface orientations suggest that dune forming winds were soutiieasterly. The
second type of dune accumulation consists of clusters of large, widely spaced straight or
curved ridges, which often intersect to create rectilinear patterns. Dunes are typically spaced
1.6-4 km apart with a mean spacing of 2.42 km. On the margins of some dunefields there are
small (0.5-0.9 km spacing) cresentic -dunes together with widely spaced barchans.
Pyramidally shaped dunes are evident in some examples and many dunes appear to have
multiple slip faces which face northwest or south. Dunefields of this type occupy a small
proportion (2-10%) of the crater floor.
Terrestrial analogues for dunes in this region which are equivalent to those found in type 1
dunefields have been discussed by Breed (1977). The dunes of the type 2 dunefields appear
to be dimensionally similar to large crescentic dunes in the Gran Desierto, Mexico and in
central Asian sand seas. Many of the large crescentic dunes in the Gran Desierto are compound
forms, with superimposed small dunes on their flanks. Others are complex forms, with
reversing or stellate dunes developed on their crests (Lancaster et al., in press). Some of the
type 2 dunefields are similar in size and arrangement of dune ridges to those in valleys of the
Mojave Desert.
Terrestrial dunes which are analogous to those in the intracrater dunefields have developed in
wind regimes in which dominant sand transporting winds are counteracted by seasonal winds
from opposed directions. Thomas (198 1) has suggested that winds in this region of Mars are
strong southeasterly in spring, but southwesterly in winter. In addition, the craters may also
generate local winds. In such situations, the existence of apparentiy complex dune forms may
be expected.
Further research is planned to examine the morphology of intracrater dunefields in more
detail and to assess their relationship to local geological settings. Wind tunnel modeling of
dune-topography interactions will help to interpret the form of dunes developed in areas of
complex topography.
264
References
Breed, C.S. 1977. Terrestrial analogues of the Hellespontus Dunes. Icarus, 30: 326-340.
Breed, C.S. Grolier, M.J. and McCauley, J.F. 1979. Morphology and distribution of
common 'sand' dunes on Mars: comparison with the Earth. Journal of Geophysical
Research, 84: 8183-8204.
Christensen, P.R. 1983. Eolian intracrater deposits on Mars: Physical properties and global
distribution. Icarus, 56: 496-518.
Cutts, J.A. and Smith, R.S.U. 1973. Eolian deposits and dunes on Mars. Journal of
Geophysical Research, 78: 4139-4154.
Lancaster, N., Greeley, R. and Christensen, P.R. in press. Dunes of the Gran Desierto,
Mexico. Earth Surface Processes and Landforms.
Peterfreund, A.R. 1985. Contemporary aeolian processes on Mars: local dust storms.
Unpublished Ph.D. dissertation. Department of Geology, Arizona State University.
Thomas, P. 1981. North-south asymmetry of eolian features in martian polar regions: analyis
based on crater-related wind markers. Icarus, 48: 76-90.
Tsoar, H., Greeley, R. and Peterfreund, A.R. 1979. The north Polar sand sea and related
wind patterns. Journal of Geophysical Research, %A: 8167-8182.
265
TIMIN6 OF FROST DEPOSITION ON MARTIAN DUNES: A CLUE TO PROPERTIES OF
DUNE PARTICLES! P. Thomas, Cornell University.
The nature of the particles in martian dunes has been a source of
controversy since the discovery of dune forms on Hars, Their color and
high thermal inertias have suggested that the dark dunes are made of medium
to coarse grained, minimally weathered basaltic or similar materials. The
common association of the dunes with the polar layered deposits, however,
has raised questions about whether the dune particles could be agglomera-
tions of much finer particles that can be carried into the polar regions in
suspension. It is also not known if the dark material at the surface of
dune fields is representative of the bulk of a dune deposit; thick dune
fields immediately adjacent to layered deposits are very suspect in this
regard. Information on the physical properties of the dunes may be avail-
able from the behavior of frost deposition and loss from the dunes. Thomas
et al . (1979) found that some dunes at high southern latitudes appeared, to
be sites of early frost deposition in the fall; James et al . (1979) noted
many dunes retained frost late into the spring. We have now mapped the
occurrences of dune brightening in the fall and have done the quantitative
photometry needed to confirm that the dunes are the sites of early frost
deposition.
Early deposition of frost on dark dunes is spectacularly displayed in
early fall images by Viking Orbiter 2, Where there are data at the appro-
priate season, virtually all the dark dunes between 58 and 72° S become
bright patches instead of dark areas a few days before the general cap edge
passes their locale. The surroundings then brighten rapidly and usually
become slightly brighter than the dunes. In the spring, some, but not all,
of the dunes remain brighter than their surroundings foT several days after
passage of the general cap edge to the south.
We have made scans across the dunes in images taken at several differ-
ent times to determine the time history of the dune albedo. We have esti-
mated atmospheric contributions using optical depth data (F. Jaquin, pens,
comm.) and the brightness of shadows in some images. Phase angles are
generally over 90° but the viewing geometry does not vary greatly so phase
functions are not a major problem. The data show that the dunes brighten
very substantially between Ls = 10° and 40°, depending on the latitude.
Bright coverings on dunes form outliers 1 to 5° north of the cap edge.
Formation of the general cap then sometimes again reverses the contrast of
the dune field with the surrounding area (Fig, 1).
Similar data are not available for the north, partly because of the
imaging sequences, and partly because of the different distribution of
dunes. However, some dunes that are outliers south of the main dunes
retain frost throughout the martian year.
Early deposition of frost on dunes, relative to surroundings, could be
caused by: (1) More rapid heat loss from dunes; (2) trapping of blown
frost; (3) volatiles in the dunes being trapped at the surface in times of
cooling atmospheric temperatures. A separate issue is whether the frost is
266
water or carbon dioxide. More rapid heat loss would imply a lower thermal
inertia than the average. Most data on dark martian dunes suggests that
they instead have a higher thermal inertia.
The dunes are trapped sediment, thus trapping of blown frost might be
expected. However^ the excellent correspondence of brightiening with the
edge of the dunes suggests the presence of dark material is the important
characteristic. Additionally, winds in the fall season appear different
from those that form the dunes (Thomas et al., 1979).
Water ice trapped in the dunes at depth and at warmer temperatures
than the cooling surface might cause deposition of water frost on the sur-
face in an otherwise yery dry hemisphere of Mars. It is not clear, how-
ever, that the small amount of water that could diffuse to the surface
could explain the amount of the albedo change or that it could help explain
the extra retention of frost in the spring. The investigation is now con-
centrating on the amounts of frost necessary to cause the albedo changes
and thermal modelling of the regolith.
This work is supported by NASA Grant NAGW-111.
References
Thomas, P., J. Veverka, and R. Campos-Marquetti (1979). Frost streaks in
the south polar cap of Mars. J.G.R . 84, 4621-4633.
James, P. B., G. Briggs, J. Barnes, A. Spruck (1979), Seasonal recession
of Mars' south polar cap as seen by Viking. J.G.R. 84, 2889-2922.
Figure: Reflectance scans across an intra-crater dune field, showing the
early fall brightening of the dunes followed by greater brightening of area
outside dunes. Data at Lg = 32° are from Viking image 510B76; at 1$ =
47°, 545B15. The I/F values have been corrected for atmospheric contribu-
tion based on nearby optical depth data, and reduced to Bof(a) with a
Minnaert k of 0.5, The dunes were 30% darker than surroundings at Lg =
353°, Part of the difference between dunes
is due to shadowing by the dunes.
and surroundings at L5 = 47^
267
WIND RIPPLES IN LOW DENSITY ATMOSPHERES
/. 5. Miller*, J.R. MarshaiHf, and R. GreeleyV '^Department of
Mathematics, University of Washington, Seattle, WA 98195, YUepartment of Geology,
Arizona State University, Tempe, AZ 85287.
The dynamic pressure of a fluid, 1/2 p U^ (p = fluid density, U = fluid velocity) is
a primary factor controlling the motion of particles transported in either a liquid or a gas.
This quantity is usually varied in experiments by making changes in U. In the Martian
Surface Wind Tunnel^ (MARSWIT), we have examined the effect of varying p on particle
transport by conducting tests at atmospheric pressures between 1 and 0.004 bar. This
study specifically concerns the effect of varying p on the character of wind ripples, and
elicits information concerning generalized ripple models ^'^ as well as specific geological
circumstances for ripple formation such as those prevailing on Mars (attn. press. = 0.0075
bar). In all the MARSWIT experiments, run times were sufficient to procure ripple
equilibrium: ripple height and wavelength X were stable with time. Tests were conducted
primarily with 95 jim quartz sand, and for each atmospheric pressure chosen, tests were
conducted at two (freestream) wind speeds: 1.1 U*^ and 1.5 U*t , where U*t is saltation
threshold.
Figure 1 shows the relationship between saltation threshold and atmospheric
pressure to be exponential. There is little variation in U*t down to ~0.25 bar, but below
this pressure, U*t rises very sharply. Ripples appeared at all pressures tested for both
wind speeds. Wavelength data is summarized in Figure 2 from which it is apparent that
there are three distinct ripple trends. Trends A and B represent relatively large "ballistic"
ripples that show a distinct increase in X with decreasing atmospheric pressure (for
threshold-normalized wind speeds). These are well defined in Figure 3 (oblique photos of
the MARSWIT test beds) for 1.5 U*^. The amplitude of these ripples dimishes with
decreasing pressure until they are no more than narrow ridges at 0.125 bar and, thereafter,
the features disappear. Trend D represents small structures (probably aerodynamic in
origin) superimposed on the larger ripples, They were best developed at 1.1 U*^ (see Fig.
3) and their wavelengths are more or less independent of atmospheric pressure. Trend C
represents laterally discontinuous ripples formed from the coarser fraction of the sand
(naturally sorted on the bed by the wind at low pressures), but their relationship to wind
speed and pressure is not very clear.
Data analysis is in a preliminary stage, but the MARSWIT data suggests: 1)
ballistic ripple wavelength is not at variance with model 2,3 predictions, 2) an atmospheric
pressure of ~0.2 bar could represent a discontinuity in ripple behavior, 3) ripple formation
on Mars may not be readily predicted by extrapolation of terrestrial observations.
References: 1) Greeley, R., Iversen, J.D., Pollack,. J.B., Udovich, N., White, B.R.
(1974), Proc. Roy. Soc. A341, 331-336. 2) Bagnold, R.A. (1941), The Physics of
Blown Sand and Desert Dunes, Chapman & Hall, 265, p. 3) Sharp, R.P. (1963), J.
Geol., 71, 517-636.
268
40
ca
9
30
<D
ft
m
•d
d
20
■p
m
<^
10
oL
0.
"C3.-^.
"•-0— ~._-.^_.____^__,
0.25
0,50
0,75
1.00
Atmospheric pressure, bar
FIGURE 1, Threshold wind speed as a function of
atmospheric pressure for 95 p^ quartz
particles
40 p
A
1.1 U*t
©
1.5 U*t
\
\
\
\
o\'
\ "^^
^
s
S B
"-.
X
"X.
C / %.
^"o*.
d
/^
""*• •»«.
0,25 0,50 0,75 1.00
Atmospheric pressure, har
FIGURE 2, Ripple wavelength as a function of
atmospheric pressure and wind speed
for 95 pm quartz particles
269
O .*!*>!
?4
45
5
•>
®
1
M
»
®
fe
ft
O
■H
i^
^ ■
a.
K
o
4^
ii^i'-^:.:-
t
i
I
.!.!>
H*<
FV-
St>i»
i20
FIGURE 3„ Rippled sand surfaces in MARSWIT produced over a range of
atmospheric pressures at two wind speeds (oblique photo-
graphs) . Wind direction is from the left; illumination
from upper left.
270
EOLIAN SALTATIOri ON HARS
D.J. MacKinnon, USGS, Flagstaff, AZ 86001
At the high wind speeds necessary for saltation on Mars, individual
sand grains assume relatively long low angle trajectories compared with
those on the Earth. Williams and Greeley [1] argued that saltating grains
on Mars do not transmit sufficient momentum to the surface for entrainment
of secondary grains, and they concluded that the Martian saltation cloud
has a low number density and the sediment flux is weak, I have developed
a computer model in order to examine the effects of both the wind field
and surface collisions on the Martian saltation cloud; results are
summarized here.
Grain trajectories were derived by numerical solution of the exact
two-dimensional equations of motion for a spherical sand grain embedded in
an arbitrary wind field [2]. The saltating grain collides with an
individual surface grain at a randomly chosen point within a collision
cross section defined by the colliding grain and the surface grains. The
surface grains lie in mutual contact on a horizontal plane. This
collision model [3] does not account for entrainment of secondary grains;
it focuses only on the primary saltating grain whose behavior dominates
the saltation cloud [4], The horizontal and vertical speeds of a
saltating grain just after collison are restored to a fraction of the
speeds before collision in proportion to two coefficients of restitution,
EN and ET (Figure 1). The vertical wind speed is zero and the horizontal
wind speed varies with height according to a logarithmic distribution [5]
as scaled by the wind-shear speed.
The horizontal speed versus incidence angle for a single saltating
grain just before each of 100 successive, random collisions with a surface
of similar-size grains is shown in Figures 1 and 2. The distribution of
incidence speeds and angles assumed by one grain in this model
characterizes reasonably well the quantitative physical processes in the
saltation cloud, Saltating grains (Figure la) that transfer small
fractions of their horizontal momentum to the surface (ET=0.5) attain
higher speeds, rise higher above the surface, and achieve longer
trajectories than do saltating grains (Figure lb) that transfer larger
amounts of momentum (ET=0.2); relatively small differences in the
coefficients of restitution yield significant differences in the saltation
cloud. These conclusions remain valid for wind-shear speeds at the
saltation threshold (Figure 2). As shown in Figure 2b, grains can both
assume long trajectories and transfer large amounts of momentum to the
surface for the rapid development of the saltation cloud. Nevertheless,
because the surface-collision parameters of saltating Martian sediments
are unknown, the development and effectiveness of the saltation cloud
remain uncertain. Surface-collision parameters for spherical sand grains
on Mars could be more accurately characterized by scaled laboratory
experiments.
References: [1] Williams, S.H., and R. Greeley, 1986, Abs. Lun. Planet.
Sci, Conf., 17th, p. 952-953; [2] White, B.R. et al , , 1975, NASA TM
X-62463, 200 p; [3] MacKinnon, D.J,, 1986, NASA TM 88383, p. 254-256,
[4] Mitha, S, et al . , 1985, Brown bag preprint series in basic and applied
sciences, Calif. Inst. Tech., BB-36, 25 p; [5] Bagnold, R., 1941, The
physics of blown sand and desert dunes: London, Chapman and Hall, 264 p
271
o
UJ
CL
iE
es
X
IV)
IV)
30.00
2S.00
20.00
IS. 00
10.00
S.00
0.00
0.00
MHRS
GRAIN DIP. =.03 CM
USTHR=100 CM/5EC
EN=0.3 ET=0.5
m H
» .4
a ajh I
« ■ B
J_
5.00
INCID.Rhffi. (DEC.)
lO.OO
ou.uu
• MRRS
1 1 . 1 .
■ GRAIN DIfl.=
.03 CM
.
2S.00
■ U5THR=100 CH/5EC
-
. EN=0.3
ET=0.2
-
a
UJ
20.00
15.00
-
sa
i
10.00
.
5.00
o.(n
■
n
. «
£."S-ii*iirllri. , >
0.00
5.00
INCID.RNG. IDEG.)
10.00
(a)
(b)
Figure 1 - Horizontal speed vs. incident angle (the angle between the
grain trajectory and the horizontal) of a saltating spherical sand grain
immediately before 100 successive, random collisions with a surface of
similar-size sand grains. Data points are derived for a wind-shear speed
(USTAR) less than the saltation threshold for 0.03-cm-diameter sand grains
on Mars. EN is the fraction of speed restored along the line of centers
between the colliding grain and the surface grain in question; ET is the
fraction restored perpendicular to the line of centers. Except for the
difference in ET, components of (a) and (b) are the same.
CO
80.00
70.00
60.00 -
SO.OO -
o
UJ
Cl.
^ _
J <->
E <n *0.00
tvl
oc
30.00
20.00 -
10.00
0.00
MRRS
GRAIN DIfl.=.03 CM
USTflR=230 CM/SEC
EN=0.3 ET=0.5
i!»
J'
j_
j_
_L
O.DO I. 00 2.00 3.00
INCID.RNG. (DEG.)
4.00
(a)
5.00
60.00
70.00
60.00
o
UJ
50. K)
UJ
Q.
in
0NTRL
(H/SEC
40.00
M
»— I
DC
IS)
30.00
X
20.00
10.00
0.00
— 1 1— 1 1 —
MRRS
GRfilN DIP. =.03 CM
USTflR=230 CM/SEC
EN=0.3 ET=0.2
.1"
i>
I
j_
0.00 1.00 2.00 3.CM
INCID.fltC. (DEG.)
4.00
(b)
Figure 2 - Same as Figure 1 except that the wind-shear speed (USTAR) is
set equal to the threshold wind-shear speed for the saltating grain.
5.00
MASS TRANSPORT BY AEOLIAN SALTATION ON EARTH, MARS AND
VENUS? THE EFFECTS OF FULL SALTATION CLOUD DEVELOPMENT
AND CHOKING.
S. H. Williams and R. Greeley, Department of Geology, Arizona State Uni-
versity, Tempe, Arizona 85287
The purpose of this study is to be able to predict the characteristics of particle
motion and tiie quantity transported by the wind under a variety of planetary environmental
conditions. Aeolian transport of surficial material is an important part of the
sedimentological cycle, and, on Earth, aeolian activity can strongly affect land utilization
and habitabiUty. Consequently, terrestrial aeolian processes have been extensively studied.
Numerous expressions for the quantity of material moved in saltation as a function of wind
conditions and particle type have been derived (see Greeley and Iversen, 1985, for review).
However, these studies do not distinguish between the contribution of particle speed and
particle concentration to the overall saltation flux. The wind speed range over which the
saltation flux predictions are valid is sensitive to particle concentration, particularly in non-
terrestrial cases. At wind speeds near saltation threshold, the predictions do not become
accurate until the saltation is fully developed, i.e., when the particle concentration (and with
it the particle flux from the surface) is high enough to fully mobilize surface material. At
sufficientiy high wind speeds, saltation is retarded when the particle concentration reaches a
value at which mid-air collisions become common, interfering with the orderly transfer of
momentum from wind to particles aloft to surface. This condition has been termed
"choking" because too much material is moving in response to the wind stress to allow
smooth flow (Williams and Greeley, 1985).
Two wind tunnels were used for this study: an open-circuit, terrestrial environment
wind tunnel (ASUWT) and a closed-circuit tunnel capable of operating at venusian
atmospheric density (for descriptions, see Greeley et al., 1985 (ASUWT) and Greeley et
al., 1984 (VWT)). Particle concentration was determined by measuring particle speed and
particle flux at tiie same height in the saltation cloud. Particle speed was measured using a
particle velocimeter modified from a USDA design (Schmidt, 1977). Particle flux was
measured using a stack of particle collectors- (see Greeley et al., 1985). The mass
concentration at the measurement height is the local mass flux, q, divided by the average
particle speed, Vp. The particle concentration at that height is the mass concentration
divided by the mass of an average particle.
The conditions necessary for full saltation development can be examined in the
wind tunnel by measuring the total saltation flux, Q, and the particle speed near the surface.
The upward particle flux from a unit surface area, Go, is Q divided by the average saltation
pathlength, L. The horizontal momentum available for transfer to the surface is Go x Vp,
per unit surface area. If gravitational differences are accounted for, surfaces composed of
similar materials should require the same momentum availability for full saltation.
Measurements taken at just above saltation threshold in the VWT for quartz particles of
250-300 ^.m diameter indicate that saltation becomes fully developed when the momentum
available is ~3 gm cm/sec per unit surface area. Similar measurements in ASUWT under
terrestrial conditions yield a value of "2 gm/cm sec. If a value of ~ 1.5 gm/cm sec is
considered, the lower value indicating the difficulty of obtaining measurements very near
saltation threshold, then the corrosponding minimum saltation flux necessary for full
saltation development on Mars is ~0.6 gm/cm sec (Williams and Greeley, 1986). Using
the saltation flux equation in White (1979), the wind speed required for full saltation is
~10% above threshold. The presence of rocks on the martian surface further disperses the
274
saltation cloud by increasing the rebound height of saltating particles. The saltation cloud,
therefore, remains poorly developed, with less material in motion than predicted by
application of equations derived for terrestrial conditions.
Choking of the saltation cloud due to high particle concentration does not occur on
Mars, but can be significant on Earth and it certainly is on Venus. The onset of choking
was determined in ttie ASUWT by observing the change in particle speed at a height of 2
cm as the wind speed increases. As the saltation cloud grows, the particle speed increases
but the ratio of particle speed to freestream wind speed decreases, due to the velocimeter
becoming relatively lower in the saltation cloud as the cloud thickens. The onset of
choking is inferred to occur when the actual particle speed at 2 cm height decreases. This
was observed in ASUWT for particles of walnut shell (density - 1.1 gm /cc) and quartz of
170-2 11 jim diameter; larger particles were not observed to display choking behavior. The
walnut shell particles showed an actual drop in speed at 2 cm height when the freestream
wind speed exceeded ~10.5 m/sec (2.0 x threshold); the quartz particles at wind speeds
above ~12.5 m/sec (2.5 x threshold). The critical particle concentrations are ~20
particles/cc for walnut shell and ~10 particles/cc for quartz. The momentum available to the
surface is about the same in each case; the difference in critical particle concentration is a
reflection of the longer pathlength of the less massive wahiut shell. Particle pathlengths on
Venus are 1-2 orders of magnitude shorter than on Earth; the "compression" of the particles
in saltation into the small volume near the surface causes choking to readily occur.
A summary of saltation behavior is as follows. On Earth, saltation fully develops at
near-threshold wind speeds and rarely chokes. On Mars, saltation becomes fully
developed only at favorable locations and at wind speeds well above threshold. The
martian saltation cloud has an insufficient particle concentration to cause choking,
particularly over a partially rocky surface. On Venus, full saltation occurs at wind speeds
very near threshold, but saltation flux will increase less rapidly than expected at relatively
modest wind speeds due to choking.
REFERENCES
Greeley, R. and J. D. Iversen, 1985, Wind as a Geologic Process, Cambridge: Cambridge University Press,
333p.
Greeley, R., J. Iversen, R. Leach, J. Marshall, B. White and S. Williams, 1984, Windblown sand on
Venus: Preliminary results of laboratory simulations, Icarus, v. 57, p. 1 12-124.
Greeley, R., S. H. Williams, B. R. White, J. B. Pollack and J. R. Marshall, 1985, Wind abrasion on Earth
and Mars, in Woldenberg, M. J. (ed.). Models in Geomorphology, Boston: Allen and Unwin, p. 373-422.
Schmidt, R. A., 1977, A system that measures blowing snow, USDA Forest Service Research Paper RM-
194, 80 p.
White, B. R., 1979, Soil transport by winds on Mars, Jour. Geophys. Res., v.84, n. B8, p. 4643-4651.
Williams, S. H. and R. Greeley, 1985, Aeolian activity on Venus: The effect of atmospheric density on
saltation flux (abs.), Lunar Plan. Sci. Conf. XVI, p. 908-909.
Williams, S. H. and R. Greeley, 1986, Wind erosion on Mars: Impairment by poor saltation cloud
development (abs.). Lunar Plan. Sci. Conf. XVII, p. 952-953.
275
Development of Wind Tunnel Techniques for the Solution of Problems In
Planetary Aeolian Processes
Robert SuiUvam' , Jeffrey Lee^, and Ronald Greeley', Departments of 'Geology
and ^Geography, Arizona State University, Tempe, AZ 85287
Aeolian processes (wind/surface interactions resulting in sediment erosion,
transport, and deposition) dominate current surface activity on Mars and are suspected to be
active on Venus (Greeley and Iversen, 1985). In the absence of returned comprehensive
meteorological data, reliable vi'ind tunnel simulations become necessary for increasing
understanding of aeolian processes on other planets. Key to determining the character of
any wind/surface interaction is U* (the wind friction speed). Obtaining U* values for the
surface of Mars, for instance, requires complete velocity-vs.-height wind profiles, but the
Viking Lander wind velocity data are only applicable to the spot elevations (1.6 m) of the
meteorology booms. An infinite number of wind profile curves can be drawn through each
1.6 m elevation. Determination of the proper wind profile requires the Viking wind
velocity data to be coupled with a knowledge of Zq, the surface roughness parameter. The
problem can be solved with sufficiently sophisticated wind tunnel simulations of the Viking
Lander sites. However, the degree to which scale model wind tunnel experiments
accurately reflect field conditions has never been tested directly. The objective of this study
is to evaluate wind tunnel experiments in predicting full-scale field results. Such a direct
comparison between wind tunnel scale models and full-scale field results can identify
working guidelines for a broad range of boundary layer geological modelling applications
on Earth, but is especially relevant and critical for the planetary context.
The field study of Kutzbach (1961) is a good candidate for a wind tunnel simulation
because the experiment is carefully specified and the results have been utilized in other
studies (e.g. Lettau, 1969). Kutzbach (1961) reports wind profiles over a series of
roughness elements on a frozen lake and how the wind profile changed as the surface
roughness was varied. The approach of the current study is to duplicate Kutzbach's
roughness arrays in the wind tunnel at 1/20 and 1/40 scales, and to compare the wind
profiles over these scale models to those derived by Kutzbach at full scale in the field.
Kutzbach measured U* and z^ for ten roughness element arrays ranging in basket density
from 1 per 48.5 m^ to 1 per 0.4 m^. Wind velocities were measured by cup anemometers
on a mast located within the basket array close to its downwind end. Depending on the
run, six to eleven cup anemometers were arrayed at heights ranging from ten to 340 cm
above the ice.
For the wind tunnel simulation, 3/8- and 3/4-inch dowel cut to the proper lengths
conveniently produces reasonably accurate 1/40 and 1/20 scale roughness element models.
The matrix of runs was as follows: 2 scales (1/40 and 1/20) X 10 arrays (specified by
Kutzbach) X 3 tunnel freestream velocities (approx. 9.3, 16.0, and 20.7 m/sec) X 2
repeats for each scale, each array, and each freestream velocity = 120 runs. In practice the
number of repeats often exceeded 2 (especially for 9 m/sec freestream runs) to improve
overall consistency of the data. The data were collected by a pitot tube rake located within
each scaled roughness array in a relative position identical to that of Kutzbach's
anemometer mast. Kutzbach, using a fixed number of roughness elements, was forced to
decrease his roughness fetch (upwind distance of the roughness array) from 80 m for the
low density runs down to 18 m for the high density runs, and discussed the possibility that
for the short fetch, high density runs the boundary layer might not have been fully
developed. This possibility poses no liability for this study, whose concern is the
simulation of real field conditions in a wind tunnel, whether real field conditions are
represented by fully developed boundary layers or not. While this represents a severe test
of reality for a wind tunnel simulation, it remains relevant, for in nature many surfaces do
not possess enough fetch to support a fully turbulent boundary layer.
276
Kutzbach reduced his field data by using the computer program of Robinson (1961,
1962), which was based on the method outlined by Lettau (1957). In this method the log-
law wind profile is assumed correct over the range of heights measured, and values of Zq,
U*, and D (zero-plane displacement) are adjusted until the sum of error squares between
the data points and the fitted log curve is a minimum. The method simultaneously finds the
best fit log curve and locates the true zero reference level.
However, the same technique is apparently not valid for the reduction of data taken
over very rough surfaces in wind tunnels. Several workers (e.g. Kawatani and Meroney,
1970; Mulheam and Finnegan, 1978; Raupach et al., 1980) have reported a transition zone
below 2-3.5 times the height of the roughness elements in which U* is not constant and the
wind log-law is invalid. The data for this study were reduced according to transition zone
restrictions specified by Raupach et al. (1980), who used roughness elements very similar
in size and shape to those of this study. The zero reference plane was taken to be the floor
of the tunnel, and the qualified data were reduced according to
Uj, = (U*70.4) In (z/Zo)
where Uj is the velocity at height z and 0.4 is the von Karman constant. Here Zq represents
merely the height at which the average wind velocity = 0, and not necessarily the "effective
roughness height" as discussed, for instance, by Greeley and Iversen (1985 pp. 42-43).
This simplification expresses the wind profile in a convenient form for future comparison.
U*' represents the modified value of U* required by the Zg simpUfication.
Before comparing the wind tunnel data of this study with the field results obtained
by Kutzbach, some comparisons can be made beween the 1/40 scale and 1/20 scale results.
iTiese comparisons are summarized in Table 1.
Table 1. Evaluation of scaling relations
1/40 scale z^ : 1/20 scale z^ ratios
predicted actual std. dev.
all velocities 1:2.0 1:3.2 1.3
Distance scaling only 9.3 m/sec 1:2.0 1:3.2 1.6
16.0m/sec 1:2.0 1:3.3 1.2
20.7 m/sec 1:2.0 1:3.3 1.7
Distance scaling with 20.6 m/sec 1/40: 1:2 - 1:3.2 1:4.8 2.9
Re scaling 9.3 m/sec 1/20
Distance scaling with 9.3 m/sec 1/40: 1:2.2 1:2.3 1.0
velocity scaling 20.7 m/sec 1/20
Length scaling alone predicts a 1:2 ratio between values of 1/40 scale Zq and 1/20
scale Zq, but the actual ratio is found to be 1:3.2. If the data are partitioned by freestream
velocity each freestream velocity subset returns essentially the same 1/40 to 1/20 scale z^
ratio : 1:3.2. Combining length scaling with Reynolds number scaling predicts a lower
1/40 scale: 1/20 scale z^ ratio, although a precise value between 1:2 and 1:3.2 is not
indicated. Reynolds numbers (and thus turbulence characteristics) are most closely
matched by comparing high freestream smaller scale runs (20.63 m/sec 1/40 scale Zq
results) with low freestream larger scale runs (9.30 m/sec 1/20 scale z^ results). However,
the actual 1/40 scale to 1/20 scale Zq ratio is found to be 1:4.8 - a relatively less accurate
prediction than for length scaling alone. The final prediction to be considered was that of
length scaling combined with velocity scaling. Velocity scaling seems appropriate,
considering that for the purposes of this study z^ has been defined as the height where the
average velocity = 0. According to velocity scaling the wind should move across the same
scale distance - the upwind fetch of an array, for example - in the same time interval at both
scales. Thus, comparison of the 1/40 scale 9.29 m/sec freesteam z^ values with the 1/20
277
scale 20.65 m/sec freestxeam z^ values should yield a 1:2.2 ratio. The actual ratio is found
to be 1:2.3. The wind tunnel simulation apparently was performed under a flow regime in
which Re scaling is dominated by the effects of simple velocity scaling. The threshold Re
below which Re scaling becomes significant was not probed for in this study.
Comparing the wind tunnel data to Kutzbach's field results proved extremely
difficult in practice. The main problem lies in differences in data reduction techniques.
Kutzbach derived his Zq and U* values for each roughness array by applying a log-law fit
over his entire height range of data. Raupach et al. (1980) and other workers have shown
this to be an invalid procedure for very rough surfaces in the wind tunnel environment.
Whether the procedure is invalid for very rough surfaces in the field environment remains
unresolved. Test examples showed that tiie two reduction techniques give widely divergent
results when applied to the same data, requiring that a single reduction technique be applied
to both raw data sets for a meaningful comparison. Unfortunately, the raw data of
Kutzbach (1961) are unavailable, and transfering the raw data from Kutzbach's published
figures proved unsatisfactory, (Data points from different runs were often difficult to tell
apart, and running even the clearer points through the original Fortran program of
Robinson (1961) gave an unsatisfactorily inaccurate reproduction of Kutzbach's results.)
The best comparison that can be made involves reconstructing Kutzbach's best-fit
profile curves (from his final results) and re-reducing a set of hypothesized data points
from each curve in a manner identic^ to that used for the wind tunnel data. Dividing the
results by 40 and by 20 provides a set of values that can be readily compared with the wind
tunnel results. (Unfortunately, Kutzbach's results for the four densest arrays were reduced
in a slightly different manner than the others, disabhng them irom the comparison.) Before
a comparison can be made, however, velocity scaling must be taken into account.
Kutzbach's reference anemometer wind speeds correspond to a scaled velocity of
approximately only 0.3 m/sec in the wind tunnel at 1/20 scale. The trend of results
suggests that if it were possible to measure values of Zg at this freestream, they might range
from 1,5 to 2.5 times the Zg values measured for the same roughness arrays at 16.0 m/sec
freestream. In any case the necessary extrapolation is somewhat extended; a factor of 1.1
times the 16.0 m/sec Zq results turns out to match Kutzbach's field results (divided by 40
and 20) the best.
Although this study suggests that wind tunnel scale models can predict the values of
important wind profile parameters measured in the field, the development of more definitive
guidelines requires a field experiment designed specifically to be compared in detail with
wind tunnel results. Such an experiment is currently in the advanced planning stages.
References
Greeley, R. and J. D. Iversen (1985) Wind as a Geological Process, Cambridge University Press, 333 pp.
Kawatani, T., and R. N. Meroney (1970) Turbulence and wind speed characteristics within a model canopy
flow field, Agr. MeteoroL, 7, pp. 143-158.
Kutzbach, J. E. (1961) Investigations of the modification of wind profiles by artificialy controlled surface
roughness in Studies of the Three-Dimensional Structure of the Planetary Boundary Layer, Ann. Rep., Dept
MeteoroL, Univ. Wisc.-Madison, pp. 71-113.
Lettau, H. H. (1957) Computation of Richardson numbers, classification of wind profiles, and determination
of roughness parameters in Exploring the Atmosphere's First Mile, vol. 1, pp. 328-336.
Lettau, H. H., (1969) Note on aerodynamic roughness-parameter estimation on the basis of roughness-element
description, Jour. Appl. MeteoroL, 8, pp. 828-832.
Mulheam, P. J. and J. J. Finnigan (1978) Turbulent flow over a very rough, random surface, Boundary-Layer
MeteoroL, 18, pp. 373-397.
Raupach, M. R., Thorn, A. S., and I. Edwards (1980) A wind-tunnel study of turbulent flow close to regularly
arrayed rough surfaces. Boundary -Layer MeteoroL, 18, pp. 373-397.
Robinson, S. M. (1961) A method for machine computation of wind profile parameters in Studies of the
Three-Dimensional Structure of the Planetary Boundary Layer, Ann. Rep., Dept. MeteoroL, Univ. Wisc.-
Madison, pp. 63-68.
Robinson, S. M. (1962) Computing wind profile parameters, /. Atmos. Sci., 19, pp. 189-190.
278
AEOLIAN ABRASION ON VENUS i PRELIMINARY RESULTS FROM
THE VENUS SIMULATOR
J.R. Marshall, R. Greeley, DM. Tucker, and JM. Pollack, Dept. of
Geology, Arizona State University, Tempe, AZ 85287
The Venus Simulator is designed for testing the mechanical effects of aeolian abrasion
on rocks and particles at the surface of Venus. The Venus Simulator will impact sand or
pebble-size particles with controlled velocity and controlled periodicity against a rock target
in a carbon dioxide atmosphere at temperatures up to 770K and pressures up to 114 bar.
These extreme conditions are achieved in the pressure vessel depicted in Fig. 1 which has
an internal volume of 0.05 m^. The vessel contains a 7 cm diameter, 75 cm long, tubular
furnace which provides an electricaUy-heated gas reservoir. An abrasion device is inserted
through the center of one of the end flanges into the reservoir and is viewed through a 5 cm
thick quartz window. Illumination of the device is through a light pipe at the opposite end
of the pressure vessel. A gas-pulsing system (Fig. la) produces rapid-cycle release of
internal pressure and, in so doing, causes gas to be drawn from the reservoir through a 2
cm-long gas gun in the abrasion device. This flow projects particles at a rock target
situated directiy in the gas stream. Pressure in the vessel is maintained by a gas-intensifier
system. The flowing gas is at the same temperature as the impactor and target.
The present test series is examining the role of atmospheric pressure on aeolian
abrasion for a constant temperature of 737K. Results from a 20 bar/737K test are depicted
in Figure 2. Both the rock target and the impactor were fine-grained basalt. The impactor
was a 3 mm diameter angular particle chosen to represent a size of material that is
entrainable by the dense venusian atmosphere and potentially abrasive by virtue of its mass.
It was projected at the target 10^ times at a velocity of 0.7 m/s (determined by high-speed
video filming). The impactor showed a weight loss of ~1.2 x lO^^ gm per impact
(calculated from the change in geometry) with the attrition (readily apparent from Fig. 2a &
b) occurring only at the edges and comers. The arrow on one of the faces of the particle
indicates a spot viewed in the SEM before and after impact and no damage had occurred.
The impactor edges developed irregularly-defined surface layers of mechanically "bruised"
material (Fig. 2c) and small impact pits with fracture patterns similar to those developed at
room temperature. However, stylus profilometry and optical microscopy step-profiling
suggested that the target had gained material, but further tests are required to substantiate
this finding. The surface texture in Figure 2d certainly has the appearance of surface
smearing of material, and there were no clearly-defined fracture patterns indicating
chipping. Weighing the target before and after the test showed weight loss but was
inconclusive because control samples of basalt subjected to the same conditions, except for
impact, also lost weight.
It is concluded from these results that particles can incur abrasion at venusian
temperatures even with very low impact velocities expected for Venus, but the impacted
rocks may present some surprises.
279
FLANGES
ELECTBICAL PORTS
FOB THERMOCOUPLES
AND HEATERS
LAMP
HOUSING
GAS-PULSE
CONTROL
VENT LINE
QUARTZ
WINDOW
VENT LINE
STAINLESS STEEL
'O' RING
SUPPORT ARM
GAS GUN
50 em
B
THERMOCOUPLE .
■ PROBE g»5-5»fSjii
VENTING ORIFICE
GAS GUN
UV//////////^//////////^ QUARTl
ILLUMINATION
HEATING
ELEMENTS
VENTING PIPE
k% -f-WINDOW ^
WINDOW
VIEWING
DIRECTION
FLANGE SEAT
SEALING FLANGE
SUPPORT ARM
-HEATING TUBE
-INSULATION
Scm
FIGURE 1. VENUS SIMULATOR
ft. Pressure vessel sectional vieH showing internal heating
arrangement and position of the gas-gun abrasion device,.
B. Abrasion device sectional view Cenlargewent of area shown in
A). Components enclosed by heavy outline form a single unit that
is inserted into the pressure vessel through the right-side
flange. The impactor rests on a wire mesh at the base of the gas
gun until a momentary gas pulse projects it upward against the
rock target. Sas flan is induced In the gun by exhausting gas
through the venting pipei gas flows from the reservoir through
the mesh and past the target. After each venting pulse, the
impactor falls back to the mesh»
280
c
$"■
i^-,'-
■■ "■.■•ii
*» '■?-'
■.**« ^
.V; .
■ f *
■ i
■^ .we I
Ja^
.si-
cV
i<T
&■'
.JW
.**
i*\
ii*x
. f
'0 ■!■:■
FIBURE 2. IMPACT TEXTURES ON TARGET AND IMPACTOR
A. Impacting particle before abrasion tests note sharp edges and rough
■fracture faces.
B. Impacting particle after abrasion tests attrition Is very apparent —
corners and edges have been subdued. Arrow at top of particle indicates
position of area shown in C. Arrow on left face of particle indicates
position Mhere impact damage is absent (see text).
C. Corner of impacted particle showing mechanical bruising teKture Ccenter
of photo) and small impact pits tupper and lower left corners of photo).
D. Impacted target surface showing highly— irregular pattern of surface
layer which may be mechanically-bruised target material, or Sffleared
comminution debris transferred from the impactor.
281
REYNOLDS NUMBER EFFECTS ON SURFACE SHEAR STRESS
PATTERNS AROUND ISOLATED HEMISPHERES
J A. Lee^ and R. Greeley^' ^Department of Geography and ^Department of
Geology, Arizona State University, Tempe, Arizona 85287
Obstacles projecting into the wind stream alter the shear stress on the surface around
them, thus altering the erosion, transportation, and deposition of aeolian sediment. This
study is concerned with the effect of Reynolds number on the pattern of shear stress on the
surface around an isolated hemisphere. An understanding of Reynolds number effects is
necessary if wind tunnel results are to be scaled up to natural situations for meaningful
applications.
Surface shear stress was measured using the naphthalene sublimation technique outlined
by Lee and Greeley (this volume). The hemisphere used is 0.033 m.high and was
immersed in the lower portion of the 0.13 m boundary layer. Surface shear stress, to, is
often presented as the friction velocity, u*, where u* = (Pa to)^*^ with pa the atmospheric
density.
For this study Reynolds number. Re, is defined as
(1) Re=(u* h)/v
where h is the hemisphere height and v is the kinematic viscosity (1.164 x lO^^ m^ sec-1
for the 35 °C air used in the experiments (from Oke, 1978)).
Figure 1 displays the surface shear stress patterns for Re of 1360 to 2977. At the
threshold of motion for fine sand this corresponds to object heights of 0.10 to 0.22 m on
Earth, 0.76 to 1.67 m on Mars, and 0.024 to 0.053 m on Venus (all values calculated from
information in Fig. 3.17 of Greeley and Iversen, 1985).
Over the entire range of Re there is an increase in relative shear stress at all locations
except for a small zone immediately downwind of the hemisphere. The highest relative
shear stress is immediately in front of the object, presumably caused by reversing flow in
the lowest layers of the boundary layer due to fluid impact on the object, (see Baker, 1979,
for a more detailed discussion of this phenomenon). The region of low relative shear stress
is due to sheltering by the object. Relative shear stress is high behind this region due to
increased turbulence, as found by Greeley et al. (1974) for raised rim craters and Greeley
(1986) for domical hills.
The most obvious change in the pattern as Reynolds number increases is in the strength
of the horseshoe vortex. Figure 1 A shows a strong vortex downwind of the edge of the
hemisphere. As Re increases, however, the horseshoe vortex decreases in strength. This
is likely to be due to an increase in flow separation with Re resulting in the flow tending to
wrap around the object at lower wind speeds and flow over it at higher speeds. In Figure
ID a reattachment zone appears from x = 2h to x = 3.5h along the centerhne.
In a discussion of air flow around objects Snyder (1972) cites evidence suggesting that
for a given object shape there is a minimum Reynolds number above which the flow
characteristics are essentially Reynolds number independent. The value of this Reynolds
number appears to increase as shapes become more streamlined. Figure 1 shows that for
hemispheres this minimum Reynolds number has not been achieved at the wind speeds
282
used. (For Reynolds number defined by frees tream wind speed, n^, this corresponds to a
range of (u<„ h)/a) =29515 to 58888.)
This experiment shows that the surface shear stress pattern is strongly affected by
Reynolds number, at least within the range of Re used. The strength of the horseshoe
vortex decreases with increasing Re. This is presumably due to a decrease in flow around
the sides of the hemisphere and an increase in flow over the object as Reynolds number
increases.
References
Baker, C.J., 1979, The Laminar Horseshoe Vortex: Journal of Fluid Mechanics, v. 95,
pt. 2, p. 347-367.
Greeley, Ronald, 1986, Aeolian Landforms laboratory Simulations and Field Studies, in,
Nickling, W.G., ed., Aeolian Geomorphology: Allen and Unwin, New York.
Greeley, R., Iversen, J.D., Pollack, J.B., Udovich, N., and White, B., 1974, Wind
Tunnel Studies of Martian Aeolian Processes: Proceedings of the Royal Society of
London, Series A, v. 341, p. 331-360.
Greeley, R., and Iversen, J.D., 1985, Wind as a Geological Process on Earth, Mars,
Venus and Titan: Cambridge University Press, Cambridge.
Lee, J.A., and Greeley, R., this volume. Determination of Surface Shear Stress with the
Naphthalene Sublimation Technique.
Oke, T. R., 1978, Boundary Layer Climates: Methuen and Co., New York.
Snyder, William H., 1972, Similarity Criteria for the Application of Fluid Models to the
Study of Air Pollution Meteorology: Boundary-Layer Meteorology, v. 3, p. 1 13-134.
283
Figure 1 . Map views of surface shear stress patterns around a hemisphere. Flow is from
negative x/h to positive x/h. Symmetry can be assumed for the patterns. Isolines show
relative friction velocity: measured u* divided by u* on undisturbed portion of naphthalene
surface. Location relative to the hemisphere center is made dimensionless by dividing x
and y by object height. '?' indicates area where pattern is not clear with the measurement
grid used.
Y/h o.
284
DETERMINATION OF SURFACE SHEAR STRESS WITH THE
NAPHTHALENE SUBLIMATION TECHNIQUE
J A. Lee^ and R. Greeley^, ^Department of Geography and ^Department of
Geology, Arizona State University, Tempe, Arizona 85287
Introduction
Aeolian entrainment and transport are functions of surface shear stress and particle
characteristics. Measuring surface shear stress is difficult, however, where logarithmic
wind profiles are not found, such as regions around large roughness elements. Presented
here is an outline of a method whereby shear stress can be mapped on the surface around
an object. The technique involves the sublimation of naphthalene (CigHg) which is a
function of surface shear stress and surface temperature.
The naphthalene sublimation technique is based on the assumption that the transfer of
momentum, heat and mass are analogous. This assumption is known as the Reynolds
analogy (see Kays and Crawford, 1980; Schlichting, 1979). If the Reynolds analogy can
be shown to be correct for a given situation, then knowledge of the diffusion of one
property allows the determination of the others. The naphthalene sublimation technique
was developed for heat transfer studies (see Eckert, 1976) and can be readily applied to the
determination of shear stress.
Analytical Framework
The naphthalene mass transfer coefficient, C^^, is defined as
(1) Cmt = qns/Pns
where q^^ is the mass transfer flux of naphthalene from the surface and p^g is the density of
naphthalene at the surface. Goldstein, et al. (1985) outline a procedure for determinining
Cjnt for naphthalene sublimation experiments.
A value of the heat transfer coefficient allows the calculation of the surface shear stress,
Xg. The Stanton number, St, is a dimensionless term defined as (Schlichting, 1979, p.
708)
(2) St = Cn,t/(PaCpU«,)
where Pa is the atmospheric density, Cp is the specific heat of naphthalene and U^o is the
freestream velocity. The local coefficient of skin friction, Cf, is also dimensionless and is
defined as (Schlichting, 1979, p. 143)
(3) C{=x,/(0.5p^Vj).
If the Reynolds analogy is acceptable, then St can also be expressed as (Schlichting, 1979,
p. 708)
(4) St = 0.5cf
Combining Equations (2), (3) and (4) and solving for x^ yields
(5) Xs = (C^tU«)/Cp
Shear stress is often expressed in terms of the friction velocity, u*, which is defined as
(6) u* = (Xspa)0-5
285
Data Acquisition
The naphthalene sublimation technique requires information on the change in surface
elevation, surface temperature of the naphthalene and freestream velocity during a wind
tunnel run. The measurement surface is obtained by casting naphthalene in an aluminum
mold. Surface heights are determined for points on a grid network before and after a wind
tunnel run using a linear variable differential transducer mounted on an X,Y positioning
device.
Surface temperature is monitored with a thermocouple embedded in the naphthalene
with its tip near the surface. Average temperature during the wind tunnel run is used in the
analysis.
The naphthalene sublimation technique was tested in the Arizona State University
Planetary Geology Wind Tunnel. The tests involved eight wind tunnel runs during which
the average sublimation depth (of twenty points) on the naphthalene surface was
determined and the wind profile was measured during each run using a boundary layer
rake.
The friction velocity was determined both from the naphthalene sublimation technique
as described above and the wind profile. The profile data were reduced following a
procedure outlined by Lettau (1957) to determine the aerodynamic roughness length, Zg,
which was used in the logarithmic wind profile equation (Greeley and Iversen, 1985, p.
42)
(7) u* = (0.4u2)/(ln(z/zo))
where u^ is the wind speed at height z
Figure 1 shows the relationship between the two techniques for determining u*. The
naphthalene sublimation technique calculates u* values one order of magnitude lower than
the wind profile technique. This discrepency can be caused by a number of factors. The
most likely are 1.) the wind profile is determined on the masonite wind tunnel floor while
the naphthalene has a smoother surface, 2.) residue from silicon spray (used in the
naphthalene casting) may reduce the sublimation rate, and 3.) the naphthalene surface
temperature caimot be measured with a thermocouple; temperature just below the surface is
measured.
The linear relationship between the friction velocity values calculated with the two
methods (correlation coefficient = 0.99) allows the naphthalene sublimation data to be
adjusted (by linear regression) to realistic values for the wind tunnel floor.
Conclusion
The naphthalene sublimation technique as ouflined here is a reasonably accurate method for
determining surface shear stress. Its most useful appUcation is in determining the spatial
variation of shear stress around objects where numerous point values are needed.
Acknowledgement
Dr. M.K. Chyu (Arizona State University Department of Mechanical and Aerorspace
Engineering) provided invaluable advice on the naphthalene sublimation technique.
286
References
Eckert, E.G.R., 1976, Analogies to Heat Transfer Processes, in, Eckert, E.G.R., and
Goldstein, R.J., eds.. Measurements in Heat Transfer, Second Edition, McGraw-Hill
Book Company, New York, p. 397-423
Goldstein, R.J., Chyu, M.K., and Hain, R.C., 1985, Measurement of Local Mass
Transfer on a Surface in the Region of the Base of a Protruding Cylinder with a
Controlled Data Acquisition System: InternationalJournal of Heat and Mass Transfer,
V. 28, p. 977-985.
Greeley, R., and Iversen, J.D., 1985, Wind as a Geological Process on Earth, Mars,
Venus and Titan:: Cambridge University Press, Cambridge.
Kays, W.M., and Crawford, M.E., 1980, Convective Heat and Mass Transfer, Second
Edition: McGraw-Hill Book Company, New York.
Lettau, H.H., 1957, Computation of Richardson Numbers, Classification of Wind
Profiles, and Determination of Roughness Parameters, in Lettau, H.H., and Davidson,
B., eds.. Exploring the Atmosphere's First Mile, v. 1: Pergamon Press, New York p.
328-336.
Schlichting, H., 1979, Boundary-Layer Theory, Seventh Edition: McGraw-Hill Book
Company, New York.
1.4
1.2
1
u*prof °-S
(m/sec) Q g ..
0.4 ■•
0.2 ■•
0.01
0.012 0.014 0.016
u*naph (m/sec)
0.018
0.02
Figure 1. Friction velocity determined from naphthalene sublimation (u* nap^) and wind
profile (u* prof) during the same wind tunnel run.
287
CHAPTER 9
FLUVIAL PROCESSES
GROUNDWATER SAPPING CHANNELS s SUMMARY OF EFFECTS OF EXPERIMENTS WITH VARIED
STRATIGRAPHY
Kochelj R« Craig and Simmons, David W. , Department of Geology,
Southern Illinois University, Carbondale, IL 62901
Experiments in our recirculating flume sapping box have modelled valley
formation by groundwater sapping processes in a number of settings. We have
examined the effects of the following parameters on sapping channel
morphology; 1) surface slope; 2) stratigraphic variations in permeability
cohesion and dip; and 3) structure - joints and dikes.
Figure 1 illustrates the variety of designs used to simulate the
gently-dipping strata and joints characteristic of the Colorado Plateau.
Run 3 can be viewed as a control because it used uncemented, homogeneous
sediment. Runs 8-11 were designed to observe the effect of joints in a
variety of stratigraphic settings. Joints were constructed by excavating
the fine sand in linear troughs and backfilling with coarser, more permeable
sand. Runs 27 and 28 investigated the effects of varying cohesion.
Cohesion was varied by mixing different amounts of cement (between 0.5 and
5% cement) or loess in the fine sand and by using sediments of varying grain
size.
Slope of the sediment surface and the slope of the internal
stratigraphy were varied between runs to determine the effect of slope upon
the rate of sapping. In particular, we were interested in how slope affects
the rate of sapping channel development and sapping channel morphology.
Initial experiments using homogeneous sediment (Run 3) indicated that there
existed a threshold slope of about 9 below which no sapping channels
formed. Below this critical slope, a seepage face formed, but channel
incision failed to occur because sediments were not entrained. This slope
value is probably diagnostic of the fine sand used in these experiments.
Experiments with initial slopes above 11 experienced significant slumping
at the expense of channel formation.
The effects of slope in experiments with layered stratigraphy appear to
be more complex. Variations in the dip of strata seem to be more important
in channel development than surface slope. Runs 14 and 27 contained layered
strata with markedly different surface and dip slopes. The 9 surface slope
of Run 14 should have resulted in rapid channel formation. However, only
small channels formed directly above the toe of the coarse layer. Slumping
occurred downslope from this point. The stratigraphy in Run 14 was parallel
to the surface slope and no layering was exposed on the seepage face. Most
of the groundwater discharge through the coarse layer either flowed along
the flume floor to induce slumping or emerged directly above the toe of the
coarse layer, taking the shortest route through the fine layer. The surface
slope of Run 27 was only 3 , well below the threshold of transport seen in
homogeneous fine sand, but experienced significant channel development.
Channels developed because the coarse, permeable layer was exposed on the
face of a low scarp at the toeslope.
291
The depth of the sapping canyons also appeared to have been directly
related to the thickness of the sediment in the upper strata* In situations
where there was a more permeable upper layer (weakly cemented) over a less
permeable base (strong cement or loess mixture), the basal layer acted as a
base level control on incision. Width of the sapping channels varied
considerably with the thickness of the strata and cohesion. Cohesion
limited the rate of lateral cutting by retarding the rate of channel wall
slumping, resulting in narrower valleys.
Laity and Malin (1985) drew attention to the role of structure in
controlling the pattern of channel networks developed by sapping in the
Navajo Sandstone of the Colorado Plateau, They noted that on a regional
scale tributaries to the Escalante River were asymmetrically distributed on
opposite sides of the channel. Structural features such as joints and
faults create zones of increased permeability in consolidated rocks which
are preferred paths for groundwater flow. Runs 8-11 were designed to
observe the effect of joints upon the development of sapping channels. In
general, main channel trends followed joint patterns and tributaries
developed parallel to joints. Our experiments with linear zones of
increased permeability suggest that if significant joints are present,
sapping valleys will preferentially extend along these avenues of increased
groundwater discharge. The degree of influence joints will have upon
channel location probably depends upon the relative differences between the
permeabilities of the joints and the host rock. As this difference becomes
greater, the influence of jointing should become more pronounced.
We are currently experimenting with the development of channels by the
combination of groundwater sapping and rainfall runoff processes. Channels
formed in Run 28 were established first by sapping and then supplementd by
periodic rainfall. The resulting channels exhibited a tapering head area
more indicative of runoff valleys and was also characterized by more
bifurcation than valleys produced only by sapping. However, close
inspection revealed a distinctive scalloped morphology of alcoves developed
along the channel walls.
We noticed the importance of groundwater piracy in the evolution of
channel networks during most of the sapping experiments. Subsurface piracy
was commonplace in all types of stratigraphic settings and regardless of the
presence of joints. Between 3 and 6 channels typically formed at regularly-
spaced positions across the seepage face during the initial few hours of
sapping runs. During the course of a run, one or two of these channels
extended headward more rapidly until it captured groundwater from
surrounding areas that would have flowed into neighboring channels. Once
dominance of a given channel began, the process became self- enhancing and
the disparity between development of neighboring channels became even more
apparent. Eventually, the pirated channels became inactive and channel
evolution was terminated.
Groundwater piracy was best developed during several runs 7 and 18
where conditions were established to mimic the intersection of channels with
292
subsurface high-level aquifers in experiments designed to simulate channel
development on Hawaii (Kochel and Piper 1985)9 Head regions of the first
channels that reach the high-level aquifers widened dramatically and grew at
the expense of neighboring channels that were pirated. This process is
probably responsible for the light-bulb shaped valleys typical of sapping
valleys on the Hawaiian Islands,
These kinds of modelling experiments are particularly good for: 1)
testing concepts; 2) developing a suite of distinctive morphologies and
morphometries indicative of sapping; 3) helping to relate process to
morphology; and 4) providing data necessary to assess the relative
importance of runoff, sapping, and mass wasting processes on channel
development. The observations from the flume systems can be used to help
interpret features observed in terrestrial and Martian settings where
sapping processes are through to have played an important role in the
development of valley networks.
References
Kochel, R,C,, Howard, A,D. , and McLane, C. , 1985, Channel networks developed
by groundwater sapping in fine-grained sediments; Analogs to some
Martian valleys; in Woldenberg, M, , ed,. Models in Geomorphology; Allen
& Unwin, Boston, p, 313-341,
Kochel, R.C, and Piper, J,F,, 1985, Morphology of large valleys in Hawaii;
Evidence for groundwater sapping and comparisons to Mars; 17th Lunar
Planet, Sci, Conf,, Houston, p, 424-425.
Laity, J.E,, and Malin, M,C,, 1985, Sapping processes and the development of
theater-headed valley networks on the Colorado Plateau: Geol, Soc, Amer,
V, 96, p, 203-217,
RUN 3
RUN 27
RUNS
RUN 9
RUN 11
RUN 14
RUN 28
Plan View - Runs 8, 9 , 1 1
Joint Structures
EXPLANATION gg
13 fine sand
m
coarse sand
& joints
coarse sand
fine sand
& loess
Cemented
weak
^ strong
SCALE
(cm)
200
100
Fig, 1, Schematic of stratigraphic styles used in sapping experiments.
293
FLUVIAL VALLEYS ON MARTIAN VOLCANOES
Victor R. Bakerj Department of Geosclences and Department of
Planetary Sciences, University of Arizona, Tucson, Arizona 85721; and
Virginia C, Gullck, Department of Geosclences, University of Arizona,
Tucson, Arizona 85721
Channels and valleys have been known on the Martian volcanoes since
their discovery by the Mariner 9 mission. Their analysis has generally
centered on Interpretation of possible origins by fluvial, lava, or
viscous flows (debris, lahar, etc.)« As summarized by Baker (1982),
fluvial and related degradational processes (sapping) produce landforms
strikingly similar to those observed on some Martian volcanoes. However,
the possible fluvial dissection of Martian volcanoes has received scant
attention in comparison to that afforded outflow, runoff, and fretted
channels (Mars Channel Working Group, 1983).
Valley and/or channel forms on Martian volcanoes have not received a
systematic investigation, and some confusion has been generated by studies
with a local perspective. For example, Milton (1973) interpreted a single
high-resolution Mariner 9 picture from the northwest flank of Alba Patera
as displaying dendritic fluvial gullies. Carr and others (1977) used
Viking imagery of the whole region to demonstrate an origin by lava tube
and channel formation. However, in a preliminary mapping study of the
Alba channels /valleys, we find that fluvial valleys may also occur with
lava channels. The fluvial valleys occur in networks with remarkably high
magnitudes for Mars (Table 1). They are well integrated and concentrated
on the northern flank of the volcano, in areas where lava flow morphology
is subdued. Nearby areas with prominent lava flow fronts and ridges show
classic lava channel and tube morphologies.
We have initiated photo interpretive, mapping, and morphometric
studies of three Martian volcanoes: Ceraunius Tholus, Hecates Tholus, and
Alba Patera. Ceraunius is drained by radial valleys. Interpreted as
fluvial in origin by Sharp and Malln (1975). Relmers and Komar (1979)
summarized evidence that the valleys (or channels) of both Ceraunius and
Hecates were not formed by lava erosion, lava tube collapse, or tectonic
fracturing. Some lava channels and collapsed tubes exist, but we have
found that, as on Alba Patera, these have distributary patterns and
discontinuous, irregular surface morphologies. Fluvial valleys, in
contrast, are continuous and display tributary development. Some have
prominent fan deposits at their mouths, indicating sediment transport and
deposition. Fans are especially prominent on the western flank of
Hecates, where valleys appear to have reached an advanced state, perhaps
by enlargement through sapping.
Preliminary morphometric results (Table I) indicate that, for these
three volcanoes, valley junction angles Increase with decreasing slope.
Drainage densities are quite variable, apparently reflecting complex
Interactions in the landscape-forming factors described above.
294
Five other Martian volcanoes are characterized by prominent valley
systems! Uranlus Tholus, Uranlus Patera, Apolllnarls Patera, Hadrlaca
Patera, and Tyrrhena. Many of these valleys, on preliminary investiga-
tion, appear to have formed or were extensively modified by lava erosional
activity. Tyrrhena has been interpreted as having a history of pyro-
clastic activity (Greeley and Spudls, 1981)e Its radial system of troughs
resembles valleys formed by sapping (Baker, 1982, p, 75-77).
Ages of the Martian volcanoes have recently been reinterpreted in
studies in progress by R.G. Strom, N. Barlow, and associates. The new
data indicate that Ceraunius Tholus, Uranlus Tholus, Apolllnarls Patera,
Tyrrhena Patera, and Hecates Tholus all date from the period of heavy
bombardment. Hadriaca Patera and Uranlus Patera date from the terminal
heavy bombardment, and Alba Patera is post heavy bombardment. This
refined dating provides a time sequence in which to evaluate the degrada-
tional forms.
An anomaly has appeared from our initial study: fluvial valleys seem
to be present on some Martian volcanoes, but not on others of the same
age. The Hawaiian analogy (Gulick and Baker, 1986) may provide some
answers here. For example, ash mantling of Hecates (Mouglnls-Mark and
others, 1982) may have contributed to drainage initiation, as it does on
Mauna Kea and Kohala in Hawaii. Volcanic surfaces characterized only by
high permeability lava flows may have persisted without fluvial dissec-
tion.
Table 1. Morphometry of Valleys/Channels on Martian Volcanoes
Drainage
Volcano
Average
*
Slope
Density
(km/km^)
Magnitude
Junction
Angles
e^+e^
Ceraunius Tholus
10-12°
0.5-1.1
5+
29
2
28
Hecates Tholus
3-5°
0.6-2.3
10+
25
7
33
Alba Patera
<I°
0.3-1,1
34+
33
7
41
Reimers and Komar (1979)
REFERENCES
Baker, V.R., 1982, The Channels of Mars;
Austin, Texas, 198 p.
University of Texas Press,
Carr, M.H,, Greeley, R. , Blasius, K.R., Guest, J.E., and Murray, J.B.,
1977, Some Martian volcanic features as viewed from the Viking
orbiters; Journal of Geophysical Research, v. 82, p. 3985-4015.
295
Greeleyj R» , and Spudls, P^D., 1981, Volcanism on Mars: Rev, Geophys.
Space Phys.s v. 19, p. 13-41 »
Gullck, V,C,, and Baker, V,R., 1986, Evolution of valley networks on
Mars: The Hawaiian analog: Geological Society of America Abstracts
with Programs, v. 18, no. 6, p. 623,
Mars Channel Working Group (V.R. Baker, chairman), 1983, Channels and
valleys on Mars: Geological Society of America Bulletin, v. 94, p.
1035-1054,
Milton, D.J., 1973, Water and processes of degradation In the Martian
landscape: Journal of Geophysical Research, v. 78, p. 4037-4047.
Mouglnls-Mark, P,J., Wilson, L. , and Head, J.W,, 1982, Explosive volcanism
on Hecates Tholus, Mars: Investigation of eruption conditions:
Journal of Geophysical Research, v. 87, p, 9890-9904.
Relmers, C.E., and Komar, P.D., 1979, Evidence for explosive volcanic
density currents on certain Martian volcanoes: Icarus, v. 39, p.
88-110.
Sharp, R.P., and Malin, M.C., 1975, Channels on Mars: Geological Society
of America, v. 86, p. 593-609.
296
VALLEY DEVELOPMENT ON HAWAIIAN VOLCANOES
Victor R. Baker, Department of Geosclences and Department of
Planetary Sciences, University of Arizona, Tucson, Arizona 85721; and
Virginia C. Gullck, Department of Geosclences, University of Arizona,
Tucson, Arizona 85721
Work in progress on Hawaiian drainage evolution (Baker, 1986; Baker
and Kochel, 1984; Kochel and Baker, in press; Gullck and Baker, 1986)
indicates an important potential for understanding drainage development on
Mars. Similar to Mars (Baker and Partridge, 1986), the Hawaiian valleys
were initiated by surface runoff, subsequently enlarged by groundwater
sapping, and eventually stabilized as aquifers were depleted.
We have used quantitative geomorphlc measurements to evaluate the
following factors in Hawaiian drainage evolution: (1) climate, (2) stream
processes (surface runoff versus groundwater sapping), and (3) time. In
comparing regions of similar climate, drainage density shows a general
increase with the age of the volcanic island (Table 1). With age and
climate held constant, sapping dominated valleys, in contrast to runoff-
dominated valleys, display the following: (1) lower drainage densities,
(2) higher ratios of valley floor width to valley height (V^ ratios), and
(3) more positive profile concavities (Table 2). Some anomalies in drain-
age density development with time were Idenltlfed: (1) unusually high
values on the windward side of Mauna Kea and Kohala, where low permeabil-
ity ash deposits mantle relatively young, high permeability basalt flows;
(2) low values on Oahu because of the dominance of sapping processes;
(3) low values on Kahoolawe, Lanal, and Nihau, probably because of rain-
shadow effects -
Studies of stream junction angles (Table 3) indicate Increasing junc-
tion angles with time on the drier leeward sides of the major Islands.
However, relatively low values occur on the dry side of Oahu. On the
windward sides of the major islands, sapping processes and associated
slope changes result in variable junction angles.
The quantitative geomorphlc studies and earlier field work (Baker,
1980, 1982) yielded Important insights for Martian geomorphology. The
Importance of ash mantling In controlling infiltration on Hawaii also
seems to apply to Mars, Some valleys, such as Kaupo and Keanae on Maul,
evolve from lava surfaces, enlarge by groundwater sapping, and later
become conduits for lava flows and lahars (Baker, 1982). Some valleys on
Martian volcanoes seem to have similar experiences of multiple flow
processes.
The Hawaiian valleys also have implications for the valley networks
of the Martian heavily cratered terrains. Baker and Partridge (1986)
found evidence for two types of valleys in this area: (a) slightly older,
more dense networks on higher, probably relict land surfaces, gnd
(b) younger, less dense networks of deeply incised valleys that seem to
have grown headward at the expense of type (a) valleys. It is hypothe-
sized that these relationships indicate an evolutionary sequence similar
297
to that observed In the Hawaiian volcanoes. High-density surface-water
ravines formed Initially because of low-permeability rock types, appro-
priate climate, adequate relief, or some combination of these factors.
With time, some valleys deepened sufficiently to tap ground-water flow in
deeper, more permeable rock types. These valleys then enlarged by head-
ward growth at the expense of the older networks. The latter were
isolated as relict, degraded components of the landscape. The entire
system ceased functioning and was "frozen" in its approximate present
configuration at the termination of the heavy bombardment on Mars.
Table 1. Drainage densities (km/km ) for Hawaiian Study Sites
Hawaii Maui
Molokai Oahu Kahoolawe Lanal
Nihau
humid 0,3-7.0 1,2-3.5 2.4-5.5 1.7-5.1
areas
dry 0.2-1,1 1.2-2.4 1.2-4.0 1,0-4.0 1.4-3,9 0.9-3.5 1.8-2.9
areas
Table 2. Properties of Drainage Networks
Drainage Density
(km/km^)
Vf Ratio
Profile
Concavity
Sapping-Dominated
Runoff -Dominated
0,7-2.0
2.1-3,2
0,8-3,2
0.1-0.8
positive
negative
Table 3, Junction Angles of Hawaiian Valleys and Channels
Volcano/ Island
Humid Study Sites
'l
'2
e^+e^
34
11
45
34
5
40
38
11
49
41
19
60
38
10
52
41
16
57
Dry Study Sites
^2 ^^^2
Mauna Loa
Mauna Kea
Kohala
Maui
Molokai
Oahu
33
14
47
44
15
58
49
17
64
37
21
57
3| , 02, and [9^+62] ^^^ mean values.
298
REFERENCES
Baker, V.R», 1980, Degradation of volcanic landforms on Mars and Earth;
N.A.S.A. Technical Memorandum 82385, p» 234-235,
Baker, V.R., 1982, The Channels of Mars; University of Texas Press,
Austin, Texas, 198 p.
Baker, V.R., 1986, Evolution of valleys dissecting volcanoes on Mars and
Earth: N.A.S.A. Tech. Memorandum 88383, p. 414-416.
Baker, V.R,, and Kochel, R.C., 1984, Valley network development by spring
sapping: Geological Society of America Abstracts with Programs, v.
16, no. 6, p. 435.
Baker, V.R,, and Partridge, J.B,, 1986, Small Martian valleys: pristine
and degraded morphology: Journal of Geophysical Research, v. 91, p.
3561-3572,
Gullck, V.C, and Baker, V.R., 1986, Evolution of valley networks on
Mars: The Hawaiian analog: Geological Society of America Abstracts
with Programs, v. 18, no, 6, p. 623.
Kochel, R.C., and Baker, V.R., in press. Groundwater sapping and the
geomorphlc development of large Hawaiian valleys, ^jn^Hlggins, C.G,,
and Coates, D,R., editors. Groundwater Geomorphology: Geological
Society of America Special Paper.
299
THE APPLICATION OF FLOW COMPETENCE EVALUATIONS TO THE
ASSESSMENT OF FLOOD-FLOW VELOCITIES AND STRESSES
Paul D. Komar, College of Oceanography, Oregon State University,
Corvallis, OR 97331 .
The concept of flow competence is generally employed to evaluate
the velocities, discharges and bottom stresses of river floods inferred
from the size of the largest sediment particles transported (Baker and
Ritter, 1975; Costa, 1983). Flow competence has become an important
tool for evaluating the hydraulics of exceptional floods on Earth,
including those which eroded the Channeled Scabland of eastern
Washington (Baker, 1973), and has potential for similar evaluations of
the floods which carved the outflow channels on Mars.
For the most part, flow-competence evaluations have been
empirical, based on data compiled from a variety of sources including
major terrestrial floods caused by natural processes or dam failures.
Costa (1 983) provides a recent compilation of this data, yielding
relationships equivalent to
x^ = 26.6 D^ •2"'
and
u„ = 57 0^-46
respectively for the flood bed stress ( x^ ) and velocity ( u^, ) as a
function of the diameter D of the maximum-size gravel or boulders
transported (units of the relationships are in the cgs system). Data
used In support of these empirical equations include diameters ranging
1 to 500 cm.
Such flow-competence relationships would appear to provide a
straight-forward assessment of flood-flow stresses and velocities
based on the maximum size of gravel and boulders transported.
300
However, a re-examination of the data base and comparisons with
measurements of selective entrainment and transport of gravel in
rivers open to question such evaluations (Komar, in press). It is found
that the competence data and empirical relationships trend counter to
those obtained for selective entrainment, indicating that the
competence evaluations are affected by varying degrees of selective
size entrainment as well as by limits to the availability of extreme
particle sizes. In many instances the empirical competence equations
greatly over-estimate the hydraulics of flood flows, and it is
suggested that the better established selective entrainment equations
be used instead for competence evaluations as well. For gravels and
coarser materials, these can be expressed as the dimensionless
relationship
e^i = 0.045 (D/D5o)-0-7
for the Shields 0^- for the entrainment of a clast of individual
diameter Dj from a deposit of mixed sizes having a D^q median
diameter. The 0.045 and -0.7 coefficients are empirical, based on
several data sets such as those of Milhous (1973) and Carling (1973).
In the application to flow-competence evaluations, Dj is the maximum
size material transported, generally much larger than D^q. The
relationship indicates that the Shields 9^j for such extreme sizes will
be reduced below the 0.045 value given by the standard Shields curve as
revised by Miller et al. (1977), a curve which applies to deposits of
uniform grain sizes. This results because the larger grains within a
deposit of mixed sizes are more exposed to the flow and have smaller
pivoting angles, factors which ease their ability to be entrained by the
flow. This can be demonstrated through analyses of the forces acting
on the grain during entrainment by pivoting, rolling or sliding, an
approach which focuses more on the physical processes than the above
purely empirical relationships. However, those derived equations
require further testing by flume and field measurements before being
applied to flow-competence evaluations. Such tests are now underway.
301
REFERENCES:
Baker, V.R. (1973) Paleohydrology and sedimentology of Lake Missoula
flooding in eastern Washington: Geol. Society of America Special
Paper 14, 79 p.
Baker, V.R., and D.F. Ritter (1975) Competence of rivers to transport
coarse bedload material: Geol. Society of America Bull., v. 86, p.
975-978.
Carling, P.A. (1983) Threshold of coarse sediment transport in broad
and narrow natural streams: Earth Surface Processes, v. 8, p. 1-18.
Costa, J.E. (1983) Paleohydraulic reconstruction of flash-flood peaks
from boulder deposits in the Colorado Front Range: Geol. Society of
America Bull., v. 94, p. 986-1004.
Komar, P.D. (in press) Selective grain entrainment by a current from a
bed of mixed sizes: A reanalysis: Jour. Sedimentary Petrology .
Komar, P.D. (in press) Selective gravel entrainment and the evaluation
of flow competence: Sedimentology .
Milhous, R.T. (1973) Sediment transport in a gravel-bottomed stream:
Unpublished Ph.D. thesis, Oregon State University, Con/allis, 232 p.
Miller, M.C., l.N. McCave, and P.D. Komar (1977) Threshold of sediment
motion in unidirectional currents: Sedimentology , v. 24, p. 507-528.
302
PO^CBLE OmGM OF SOME CHANJffilS ON MMk PATEE&. MM^
S. E. Postawko and P. Mouginis-Mark, Planetary Geosciences Div.,
Hawaii Inst. Geophysics, Univ. Hawaii, Honolulu, HI 96822
Several alternative models have been proposed for the origin and mode of formation
of channels and valley networks on martian volcanoes, notably Hecates Tholus, Ceraunius
Tholus, Alba Patera. Early interpretations of Mariner 9 and Viking images suggested that
these features on Alba were lava channels (Carr et al., 1977), while those on Ceraunius
Tholus were interpreted as fluvial (Sharp and Malin, 1975) or volcanic debris channels
(Reimers and Komar, 1979). Subsequent mapping of Tyrrehna Patera (Greeley and Spudis,
1981) and Hecates Tholus (Mouginis-Mark et al., 1982) has suggested that pyroclastic
activity may have characterized eruptions on these volcanoes, and that at least for
Hecates the channels were probably formed by fluvial erosion of unconsolidated ash depo-
sits on the flanks of the volcano. As part of a continuing program to better understand the
eruptive history of the young volcanic centers on Mars, we have identified numerous chan-
nels on the flanks of Alba Patera that resemble the channels on Hecates. As a result, we
are exploring the possibility that some of the small channels on the flanks of Alba Patera
may be fluvial in origin, and are examining potential water sources and modes of forma-
tion.
There are several ways in which these Alba channels could have been formed by
fluvial action. One way is by direct rainfall, with the water coming either from degassing
of the magma, or being driven out of the regolith due to volcanic heating. Alternatively,
this water may have been deposited within the shallow regolith as ice or snow and then
melted by intrusives to form channels.
The exposed portion of Alba Patera was evidently formed at a time when any early,
dense, warm atmosphere on Mars had disappeared (if such an atmosphere ever did indeed
exist) (Cattermole, 1986a). Therefore, it is inferred that the channels formed under
climatic conditions similar to those at present-day Alba, with an atmospheric surface
pressure of ~6mb and annual average surface temperatures between 180 - 220K (Kiefler
et al., 1977). Very little water vapor (a mass mixing ratio on the order of 10^*) is required
to saturate the atmosphere at this pressure and temperature range, and even less at
greater altitudes in the atmosphere. Since the atmosphere of Mars is, in general, close to
saturation on a daily basis (Davies, 1979), the addition of water vapor from degassing
magma or evaporation/sublimation of subsurface water/ice due to intrusive heating may
well have resulted in precipitation in the vicinity of Alba Patera. This depends, in part, on
the release rate of water vapor to the atmosphere, the temperature of the gas, and how
the gas is distributed through the atmosphere. Assessing the likelihood of precipitation,
of any kind, playing a role in the formation of some of the channels seen on the flanks of
Alba Patera is a primary goal of our current research.
Studies of lava flows on Alba Patera indicate the production of large volumes of lava
during a given eruption (Cattermole, 1986b; Pieri et al., 1986); meaning either very high
rates of effusion for a relatively short period of time, or eruptions of long duration with
lower effusion rates. In either case, if the magma initially contained water, then large
volumes of lava may have exsolved a sufficient amount of water vapor (Greeley, 1986) to
supersaturate the atmosphere. However, Wilson and Head (1983) have shown that even a
very low (~0.01 wt.%) volatile content in martian magmas is sufficient to disrupt the
magma and result in explosive eruptions. As yet, no evidence of explosive volcanism on
Alba Patera has been detected. This does not eliminate the possibility of very low volatile
magmas releasing some water to the atmosphere, although whether these could provide
303
enough to carve channels is still under investigation.
Another possible source of water is the martian regolith. Calculations by Fanale et aZ.
(1986) have shown that subsurface ice is stable at the latitude of Alba Patera, even for
present-day climatic conditions. Heating in the vicinity of the volcano may have been
enough to drive substantial amounts of water out of the ground and into the atmosphere.
Calculations are presently under way to determine probable amounts and rate of
release of water by magmatic outgassing of low volatile magmas, and by warming of the
regolith. To date, we have examined the effects of volcanic gases (COg, HgO, and SOg) on
the paleoclimate of Mars, and their ability to warm the atmosphere a suflTicient amount to
permit surface flow of water or brmes (Postawko and Kuhn, 1986). From these calcula-
tions, it does not appeap that pure water cotild be released to flow on the martian surface
unless local heating by intrusives raised the local temperature of the regolith. Currently,
our efforts are therefore focused on the distribution of the vapor through the atmosphere
during and immediately after a volcanic eruption. Water released within a volcanic cloud
will likely rise quite high in the atmosphere (Wilson and Head, 1983), and may be
dispersed before saturation occurs. It is also possible that precipitation may originate
from so high up that snow/rain would sublimate/evaporate before reaching the ground.
Water released from the regolith may more easily saturate the lower atmosphere, and
thus be a more likely source for any water which may have cut channels.
REFEaSENCES: M.H. Carr et ai. (1977). /. Geophys. Res., vol. 88. p. 3985-4015. P. Catter-
mole (1986a). Lunar Planet. Sci. XVII, p. 107-108. P. Cattermole (1986b). Lwrutr PLa.net.
SciXWI, p. 105-106. D. Davies (1979), /. Geophys. Pes., vol. 84, p. 8335-8340. F. Fanale
et al. (1986), Icarus, vol. 67, 1-18. R. Greeley (1986). Symp. on MECA LPI Chntr. 599, p.
26-28. R. Greeley and P. Spudis (1981). Revs. Geophys. & Space Phys. vol. 13, p. 13-41.
H. Kiefler et al. (1977), /. Geophys. Res., vol. 88, 4249-4291. KB. Krauskopf (1979), Intro-
duction to Geochernistry UcGtbm-YLUI, 'blew York. P.J. Mouginis-Mark ei DtZ. (1982). J. Geo-
phys. Res., vol. 87. p. 9890-9904. D. Fieri et al. (1986). Rpts. Han. Geol. Prog. 1985.
NASA TM 88388, p. 318-319. S. Postawko and W.R. Kuhn (1986). Proc. Lunar Hanet. Sci.
Conf. 16th, J. Geophys. Res., vol. 91, p. D431-D438. C.E. Reimers and P.D. Komar (1979).
Icarus, vol. 39, p. 88-110. R.P. Sharp and M.C. Malin (1975). Geol. Sac. Mner. Bull., vol.
86, p. 593-609. L Wilson and J.W. Head (1983), Nature, vol. 308, p. 663-669.
304
NON-EQUILIBRIUM FREEZING OF WATER-ICE IN SANDY BASALTIC REGOLITHS AND
IMPLICATIONS FOR FLUIDIZED DEBRIS FLOWS ON MARS. J. L. Gooding, SN2/
Planetary Materials Branch, NASA/Johnson Space Center, Houston, TX 77058
Introduction . Many geomorphic features on Mars have been attributed to
Earth- analogous , cold-climate processes involving movement of water- or
ice -lubricated debris. Clearly, knowledge of the behavior of water in
regolith materials under Martian conditions is essential to understanding
the postulated geomorphic processes. Pertinent laboratory data have been
reported by D. M. Anderson and collaborators [e.g., 1-2]) but have been
based mostly on experiments with ultrafine- grained samples of clay minerals
or their admixtures. New experiments have been performed with sand-sized
samples of natural "basaltic" regoliths in order to further elucidate how
water/regolith interactions depend upon grain size and mineralogy. The new
data reveal important contrasts with data for clay-mineral substrates and
suggest that the microphysics of water/mineral interactions might affect
Martian geomorphic processes in ways that have not been fully appreciated.
Ice -Formation Experiments . Sand- and silt- sized fractions of two soils
from the stimmit region of Mauna Kea, Hawaii [3] were used as Mars -analogous
regolith materials. Makanaka glacial outwash consisted of relatively
little -weathered lithic and mineral fragments of basalts whereas Puu
Poliahu weathered tephra consisted largely of palagonitized pyroclastic
debris [3,4]. Using previously described equipment and methods [5],
differential scanning calorimetry (DSC) was used to measure temperatures of
water/ice phase transitions as wet slurries of individual soil fractions
were cooled or heated at controlled rates under a carbon dioxide
atmosphere. Freezing and melting of ice was studied as a function of
water/soil mass ratio, soil particle size, and thermal-cycle rate.
Comparison tests were done under the same conditions with U. S. Geological
Survey standard rock powders PCC-1 (peridotite) and BHVO-1 (basalt), and
with powdered Lithology A (olivine-pyroxene-maskelynite) of the shergottite
meteorite, EETA79001.
Freezing and melting temperatures of water-ice in the soils were only
weakly dependent on particle size over the silt- and sand-sized intervals.
Freezing points were essentially independent of water/soil mass ratio but
varied inversely with cooling rate. Melting points varied directly with
heating rate (Fig. 1). For both the soils and the comparison rocks, though,
freezing occurred at temperatures > 6 K below the equilibrium freezing
point of pure bulk water (Fig. 2). Depressed freezing points were not
controlled by salts in the samples (salt contents were negligible) but were
a consequence of the difficulty with which water-ice nucleates on igneous
minerals and noncrystalline weathering products. Olivine, pyroxene,
plagioclase, and glass (mafic or felsic) are all poor nucleators of
water-ice [5] . Although crystalline clay minerals are good nucleators of
water-ice [1,5], the mineraloids palagonite and allophane are demonstrably
poor nucleators [5]. Therefore, even at slow cooling rates, water mixed
with either fresh or palagonitized basalts or closely related materials can
survive metastably at temperatures significantly below 273 K.
Implications for Martian Geomorphology . Lifetimes of outburst floods with
low sediment/water ratios would probably be controlled by discharge rates
and degree of cover by ice bridges [6,7], rather than by the microphysical
phenomena reported here. However, in any process that might form a
water-based slurry (high sediment/water ratio) on Mars , undercooling of
water should be an important factor in determining the longevity and
efficacy of the slurry as an agent of geomorphic change. Degree of
305
undercooling should be a strong function of the mineralogy of the sediment.
Debris flows composed of fresh igneous materials or poorly crystalline
weathering products (e.g., palagonite) should support greater degrees of
undercooling and, therefore, greater dynamic lifetimes, than should debris
flows composed of crystalline weathering products such as smectites .
Photogeologic studies of debris flows on Mars might profit from greater
attention to possible correlations between styles and distances of
debris -flow movements and the lithologic characters of local regolith
materials.
MAUNA KEA, HAWAII GLACIAL OUTWASH
(1 25-250 A/m) UNDER CO2
WATER:SOIL = 1:1
1= 268
-
1 1
M.O.
(PURE WATER) -
-
MELTING
ONSET
M.I.
(PURE WATER)
-
--^ '
MELTING
INITIATION
-
,^_^^
FREEZING
-
^^"^-
ONSET
-
1 1
\
1
y 260
SUBSTRATES COOLED 5 K/MIN UNDER CO2
WATER: SUBSTRATE = 1:1
EQUILIBRIUM
FREEZING
MAUNA KEA, HAWAII
SUMMIT SOILS
WEATHERED
TEPHRA ,
9 GLACIAL
OUTWASH
BASALT
BHVO-1
SHERGOTTITE
EETA79001
(LITH-A)
PURE
WATER
(ON Al)
PERIDOTITE
PCC-1
0.5 1 5
THERMAL-CYCLE RATE (K/MIN)
Figure 2. Freezing temperatures for
wet slurries of silt-sized
"basaltic" materials. Error
Figure 1. Phase -transition tempera-
tures of water, as a
function of heating and
cooling rate, in slurries
prepared from Mauna Kea
glacial outwash soil.
bars represent
deviations of
measurements .
standard
replicate
References :
[1] Anderson D. M. (1968) Israel J^ Chem. . 6,
[2] Banin A. and D. M. Anderson (1975) Nature,
[3] Japp J. M. and J. L. Gooding (1980) Rept.
NASA Tech. Memo. 82385, 212-214.
[4] Porter S. C. (1979) Geol. Soc. Amer. Bull,
349-355.
255 . 261-262.
Planet. Geol. Program -
, Part II. 90, 908-1093.
1980,
[5] Gooding J. L. (1986) Icarus . 66, 56-74.
[6] Carr M. H. (1979) J^ Geophvs . Res. . 84, 2995-3007.
[7] Wallace D. and C. Sagan (1979) Icarus . 39, 385-400,
306
VOLATILE RESERVOIRS BELOW THE SURFACE OF THE ELYSIUM REGION
OF MARSs GEOMORPHIC EVIDENCE
Eric H Christiansen, Brigham Young University, Provo, Utah, 84502
Jennifer A. Hopler, University of Iowa, Iowa City, Iowa 52242.
The Elysium volcanic province contains a variety of geomorphic
evidence for the existence of large volatile reservoirs of subsurface
volatiles. Study of these landforms yields insight into the distribution
and size of these reservoirs and how they interact with the surface
environment and will ultimately place constraints on the geometry,
constitution, origin, time of formation, and temporal evolution of these
important components of the martian crust. Three principal types of
landforms appear to be related to subsurface volatile reservoirs in the
Elysium region of Mars : 1) small outflow channels; 2) large lahars; and
3) vast expanses of knobby terranes around the margins of the Elysium
dome.
Outflow Channels.
The most obvious expressions of the presence of a subsurface
volatile reservoir in this region are two relatively small outflow
channels (1). Located southwest of Elysium Mons, both channel systems
arise and cut across a broad expanse of older plains dotted by irregular
mesas and smaller knobs (knobby plains).
The anastomose Hebrus Valles system of channels is 250 km long and
emerges full -strength from an elongate depression. The source depression
is 10 km across and has narrow finger-like projections. Individual
sinuous channels are less than 100 m deep and about 1 km wide; a braided
reach is about 10 km wide. Streamlined bedforms are abundant in the
middle reach. The channels become narrower and shallower downslope.
Hebrus Valles terminate as a series of narrow distributaries. No
sedimentary deposits are obviously related to the development of the
channels. Hebrus Valles are similar to other small martian outflow
channels and appear to result from fluvial erosion following the outbreak
of a confined aquifer.
Hephaestus Fossae are a connected series of linear valley segments
which branch and cross downslope but have high junction angles. Locally,
the valley pattern is polygonal. Hephaestus Fossae are parallel to
Hebrus Valles but are considerably deeper and longer (600 km). The
rectilinear pattern of the valleys has suggested to some that the fossae
are tectonic in origin. However, unlike graben systems, Hephaestus
Fossae originate in an isolated depression similar to the source of
Hebrus Valles. Two sinuous, apparently fluvial, channels also arise from
this depression. We suggest that, Hephaestus Fossae are also of fluvial
origin and resulted from catastrophic flooding and draining of water from
beneath the surface. Hephaestus Fossae channels appear to have cut
through the knobby plains unit which overlies polygonal terrane like
that exposed at the NW end of the fossae in Adamus Labyrinthus (4).
Downcutting to, or subsurface flow at this unconformity produced a
channel pattern that was controlled by the polygonal troughs beneath the
younger knobby plains materials. Hebrus Valles channels did not excavate
this deposit and hence show more typical outflow features.
307
Mega-lahars.
Photogeologic studies of the Elysium volcanic province provide
examples of the interaction of magmatism and subsurface volatile
reservoirs to produce distinctive landforms (2, 5). Three sets of
volcanic debris flows or lahars issue from the northwest- trending system
of fractures that localized the three major volcanic constructs in the
Elysium province. The deposits are lobate in plan and have steep
well-defined snouts. Evidence that these mass flow deposits were wet
slurries and not lava or ash flows includes: 1) the intimate association
of channels with their surfaces--these channels are sinuous, form
anastomose distributary patterns, and have streamlined features on their
floors. These characteristics are consistent with the flow of water
across the deposits. 2) discrete channels issue from the base of the
lobate masses suggesting draining of water from initially wet sediments;
3) short reticulate systems of sinuous valleys cut portions of the
deposits' margins and look like seepage channels (3); and 4) numerous
irregular depressions mark other areas of the flows and have clearly
developed from a formerly smooth and more extensive deposit. These pits
may be created by the removal of volatiles by sublimation or seepage.
We have postulated that the lahars resulted from the melting of
ground ice and liquefaction of subsurface materials (2). The Elysium
volcanoes are the most reasonable sources of heat. This is consistent
with the stratigraphic evidence that lavas and lahars were nearly
contemporaneous. The contact of magma with liquid water may have resulted
in hydromagmatic explosions which can produce large quantities of easily
mobilized fine-grained material (7). The intersection of this fluid
reservoir with the regional fracture system led to the rapid expulsion of
a muddy slurry down the steep western slope of the province. These
sedimentary deposits extend nearly 1000 km down the regional slope to the
northwest and cover 10° km^. The deposits are less than 200 m thick near
their sources and are probably much thinner on average. The total volume
of the lahars may then be approximately 10^ km^. Taking a value of 30%
water by volume--a figure typical of terrestrial lahars and non-volcanic
debris flows (5)--implies that over 10^ km'^ of water were involved.
Knobby Terranes.
Knobby terrane provinces have relatively smooth surfaces with
variable proportions of knobs and flat-topped mesas. Broadly similar
knobby terranes cover approximately 3 million km^ in the Elysium region.
The knobs and mesas appear to be erosional remnants of a formerly thicker
deposit. The polygonal terrane of Adamus Labyrinthus underlies the knobby
terrane in the Amenthes quadrangle. In southern Amenthes quadrangle, the
knobby terranes have developed at the expense of an extensive plateau
marked by irregular depressions and pits. Layering is visible in the
walls of these ragged depressions. Erosional stripping of the knobby
deposit has exhumed large impact craters. North of the volcano Hecates
Tholus, knobby terranes developed at the expense of lava plains that
partially bury the undegraded precursor of the knobby terranes. Here,
large lava-capped blocks give way to smaller mesas which grade northward
into smaller knobs. Even farther north the knobby terranes disappear and
reveal underlying polygonal terrane. The knobby terranes' precursor
308
appears to have developed in middle martian history. It overlies the
polygonal terranes of Adamus Labyrinthus which are post-Lunae Planum in
age (4) and is in turn buried by Elysium lavas and lahars. The knobby
plains are also cut by the two large outflow channels noted above and
numerous small seepage channels on the western flanks of the Elysium
dome. However, evidence for fluvial erosion is not extensive and the
volume missing from the knobby plains precursor must have been either
stripped away by eolian processes or it may represent the sublimation of
water that had been sequestered in the layered deposits. The spatial
coincidence of the knobby plains with other water-related landforms lends
credence to the latter hypothesis. The degradation of the knobby plains
precursor appears to have occurred mostly before Elysium volcanism
because vast tracts of smooth lava plains bury knobby terrane; but at
least in the small region north of Hecates, knob development appears to
have persisted until the later stages of Elysium volcanism. Assuming
that most of the missing volume represents removal of volatiles, and
ignoring the extent of the knobby plains that must underlie the Elysium
volcanic province, the amount of water lost from this region may be
approximately 10^ km-^.
Implications for Sub-Surface Volatile Reservoirs at the Surface of Mars.
The evidence provided by these landforms is internally consistent
with the presence of a large relatively shallow volatile reservoir in the
Elysium region of Mars. If the geologic features described above are
reliable indicators of subsurface volatiles, they imply that:
- volatile reservoirs lie relatively close to the surface and underlie
millions of km^ in this region.
- their is no apparent latitudinal variation in the depth or thickness
of the volatile reservoirs.
- the precursors of the knobby terranes are or were important volatile
reservoirs.
- volatiles may be lost in a variety of ways from these reservoirs.
- volatiles were incorporated in an easily eroded surficial deposit
in the middle history of Mars. The ultimate origin of the water in this
reservoir is uncertain. A model to explain the preferential entrapment
of volatiles into the region's surface materials may be required.
References .
(1) Christiansen, EH (1985) Geol Soc America Abstr Prog. 17, p. 545.
(2) Christiansen, EH, and Ryan, MP (1984) Geol Soc America Abstr
Prog . 16, p. 470.
(3) Higgins, CG (1984) in RG LaFleur, Groundwater as a geomorphic
agent , Allen and Unwin, p. 18-58.
(4) McGill, GE (1986) Geophvs Res Lett 13, p. 705-708.
(5) Mouginis-Mark, PJ (1985) Icarus 64, p. 265-284.
(6) Pierson, TC (1985) Geol Soc America Bull 96, p. 1056-1069.
(7) Sheridan, MP, and Wohletz, KH (1983) J Volcanol Geotherm Res 17, p.
1-29.
309
EVIDEiSICE FOR GLACIATION IN ELYSIUM
Duwayne M. Anderson
Texas ASM University
Evidence for the existence of permafrost and the
surface modification due to frost effects and the presence
of ice on Mars dates from early observations by Otterman and
Bronner (1) ? Leighton and Murray (2) ^ Anderson et al, (3)?
Baranov (4) and Anderson and Gatto (5), The Viking landers
I and II confirmed the presence of water in the regolith and
the periodic occurrance of frost at the surface of Mars,
Bieman et al, (6) r Anderson (7). Observations from the
Viking Orbiters I and II demonstrated the presence of
atmospheric water at various concentrations in the
atmosphere of the planet (Parmer et al, [8]) and also
provided a means of documenting the accumulation and
sublimation of frost and ice in the two polar caps and at
several other locations on the planet. Later analysis of
the Viking Orbiter imagery produced evidence suggesting the
former presence of ice sheets that could have played a part
in shaping the surface of Mars (Lucchitta et al. 19]),
Similarities have been pointed out between a number of
streamlined Martian channel features and similar streamlined
landforms created by antarctic ice sheet movements,
A study of Viking Orbiter imagery of Granicus Valles
and the surrounding terrain in Elysium has produced further
evidence of glaciation on Mars, Granicus Valles is located
in the Northern hemisphere between 25 degrees and 40 degrees
latitude and 220 degrees to 240 degrees longitude. It
extends from the Elysium fossae near Elysium Mons to Utopia
planitiar the plains where Viking Lander II is located,
Granicus Valles has been characterized as an outflow channel
(Malin [10]^ Carr [11]). It is approximately 1100 km long?
about 10 km wide at the source, with distributory branches
that cover an area approximately 300 km wide. Three
distinctly different geological units are involved
(Mouginis-Mark [12], Christiansen and Greeley, [13]), Part
of the region is characterized as a Complex Vent Area, It
is one of the main volcanic centers in Elysium, Another has
been referred to as Modified Lava Plain by Christiansen and
Greeley (13) , and Compound Lava Plains and Erosional Plains
by Mouginis-Mark (12) , A portion of it was earlier referred
to as Channeled Plain by Christiansen and Greeley (13) ,
Volcanism has played an important role in developing
the landscapes of the Elysium region. Two features that
strongly resemble terrestrial moberg ridges have been found.
These features are ridge shaped, serrated mountains, very
similar to the moberg ridges described by Allen (14) ,
Terrestrial moberg ridges form as a result of subglacial,
fissure eruptions. The only apparent difference between
those observed on earth and these two Martian counterparts
is scale? the Martian features are much larger. Because the
310
size of a moberg ridge is limited by the thickness of ice
above the erupting fissure lava^ the greater height of the
Martian ridges implies a thicker ice sheet. The height of
the moberg ridge near Granicus Valles is estimated to be
about 2»6 km« The ground surface surrounding both Martian
moberg ridges appears to have two distinct levelsi it is
lowest at the base of the ridges,, rising to a level near the
top of the ridges at a distance* A possible explanation is
that subsidence occurred during formation of the Martian
moberg ridges due to the melting of ground ice near the
eruption area while at a distance most of the ground ice in
the permafrost is still present and the original elevation
has been preserved, Meltwater during and following
eruptions might have been suddenly released during
subglacial volcanism into Granicus Valles in one case and
into Hrad Valles in the other. Pluvial erosion thus could
have played a role in shaping both.
Crater size-frequency plots indicate differences in the
ages of surfaces from one region to another ^ ranging from
one to three billion years. The presence of a thick ice
sheet together with underlying permafrost could help to
explain this, A thick mantle of ice would have intercepted
and contained many meteoric impacts^ shielding the
underlying regolith. Subsequent removal of this ice by
sublimation or melting would then expose a relatively
unmarked surface to subsequent meteoric bombardment. These
surfaces^ from crater size-frequency plots^ would always
appear younger than the surrounding^ continuously exposed
terrains.
Many of the volcanic edifices in Elysium appear to have
uncommonly steep slopes at their bases. Many are quite
large and probably stood above the hypothetical r pre-
existing ice sheet. Volcanic flows and ejecta cascading
down slope would have come to rest on the surrounding ice.
Later disappearance of the ice would result in the
subsidence and redeposition of these materials, helping to
explain the abnormally steep basal slopes now evident.
References
1, Ottermanr J. and Bronner? P,E, (1966) r Martian wave of
Darkening; A Frost Phenomenon?^ Science - 153 r pp,
56-59,
2, Leighton, R, B, and Murray^ B, C, (1966), Behavior of
Carbon Dioxide and Other Volatiles on Mars^
Science . 153, pp. 136-144,
3, Anderson, D, M, , Gaffney, E, S, , and Low, P, F, (1967),
Frost Phenomena on Mars, Science . 155, pp, 319-322,
4, Baranov, I, Ya, (1959), Geographical Distribution of
Seasonally Frozen Ground and Permafrost,
311
Principles of Geocrynologyr Part 1. V. A,
Obruchev Institute of Permafrost Studiesj, Academy
of Science? U,S,S»R,
5* Anderson? D, M« ^ Gatto^ L. W, r and Ugolini? F. (1973).
An Examination of Mariner 6 and 7 Imagery for
Evidence of Permafrost Terrain on Mars,
PERMAFROSTi The North American Contribution to the
Second International Conference, Yakutsk^ Siberia,
National Academy of Sciences^ pp, 499-508,
6, Bieman? K, ? et al, (1976). Search for Organic and
Volatile Inorganic Compounds in Two Surface
Samples From the Chryse Planitia Region of Mars,
Science . 194, pp,72-76.
7, Anderson, D, M, (1978), Water in the Martian Regolith,
Comparative PI anetology . Academic Press, Inc, , pp,
219-224,
8, Farmer, C, B, , et al. (1977), Marss Water Vapor
Observations From the Viking Orbiters, J. Geophys.
Res. . 82, pp, 4225-4248,
9, Lucchitta, B, K, , Anderson, D. M. , and Shoji, H.
(1981) . Did Ice Streams Carve Martian Outflow
Channels? Proc. Third Coll, on Planetary Water,
Niagara Falls, New York, Nature , 290/5809, pp. 759-
763,
10, Malin, M, C, (1976), Age of Martian Channels. iL.
Geophys. Res, . 81, pp. 4825-4845.
11, Carr, M. H. (1981). The Surface of Mars . Yale
University Press, New Haven.
12, Mouginis-Mark, P, J,, et al, (1984). Elysium Planitia,
Mars I Regional Geology, Volcanology, and Evidence
for Volcano-Ground Ice Interactions, Earth. Moon .
Planets . 30, pp, 149-173,
13, Christiansen, E, H, , and Greeley, R. (1981). Mega-
lahars(?) in the Elysium Region, Mars, (abstract).
Lunar Planet. Sci. XII . pp. 138-140,
14, Allen, C. C. (1979), Volcano-Ice Interactions on Mars.
J. Geop h ys. Res. . 84, pp. 8048-8059,
312
WATER AND ICE ON MARS: EVIDENCE FROM VALLES MARINERIS*
B. K, Lucchittas U,S, Geological Survey, Flagstaff, Arizona 86001
An important contribution to the volatile history of Mars comes from a
detailed study of Valles Marineris, where excellent stereoimages and a
three-dimensional view of the upper Martian crust permit unusual
insights: the subsurface in the equatorial region of Mars below about 1 km
depth was not desiccated until relatively recently, even though desiccation
is predicted by models of the equilibrium between water in the ground and
in the atmosphere [1].
The evidence that ground water and ice existed until relatively
recently or still exist in the equatorial area comes from observations of
landslides [2,3], wall rock [4], and dark volcanic vents [5,6]. Several
observations suggest that landslides were lubricated by water. Three young
slides generated an outwash fan and gave rise to a channel that has several
bends and extends on a gradient of 4 m/km for a total distance of 250 km
from its source [7]. Also, the material in this channel was capable of
erosion at considerable distance from its source; it breached a bedrock
ridge, carved flutes in the lower channel, and eroded its banks. Doughnut-
shaped hills within this channel resemble moraines containing kettle holes,
which on Earth are formed by the melting of blocks of ice.
Some landslides have lobes that angle backward from the main debris
mass and flow downhill, others give rise to small sinuous valleys, and many
small landslides are surrounded by levees like terrestrial mudflows. These
observations also suggest that the landslide deposits contained fluids. A
small channel debouches from a tributary canyon to Valles Marineris;
apparently water discharged from the canyon walls, if canyon tributaries
were indeed formed by sapping [9],
Valles Marineris landslides are different in efficiency from large
catastrophic landslides on Earth, Whereas terrestrial landslides increase
in efficiency (distance traveled) with increasing weight [9], the large
landslides in Valles Marineris retain the same efficiency regardless of
weight [3], One explanation for the difference might be that the Martian
slides are lubricated by water, whereas most large terrestrial slides are
dry-rock avalanches [10],
A comparison of landslide speeds also suggests that the Martian slides
contained water, /^ong large catastrophic landslides on Earth, only the
Huascaran slide [11] matches the Martian ones in speed [3]; the Huascaran
slide contained much water and ice. Because all landslides in Valles
Marineris are released from wall rock, some layers within the walls that
are 7-10 km high must have contained these lubricating materials.
A relatively young, level deposit embaying eroded layered beds occurs
in the lowest area of the central troughs of Valles Marineris [12]. The
deposit looks like a dry lake bed or alluvial flat, which suggests that wet
debris contributed to its formation. The wet debris was apparently derived
from landslides or wall rock.
That Valles Marineris wall rock contained water or ice is further
suggested by its difference from the interior layered deposits. Landslides
having flow lobes that extend far out onto the chasma floors debouch only
from wall rock or erosional remnants of wall rock. No such landslides come
from the layered deposits, even where the layered deposits are as high and
steep as the wall rock. Apparently, landslides formed from the wall
materials flowed easily; those from the interior deposits generally did not
[13],
313
Faults and fault zones in Valles Marineris also shed light on the
problem of water content in the walls. Contrary to what is commonly seen
on Earths many fault zones in Valles Marineris are more resistant to
erosion than the country rock. Spurs projecting into Valles Marineris
developed along faults [4]^ and all the median ridges of wall rock
paralleling chasma walls or separating chasmata from each other occur where
faults and fracture zones are densely spaced. Apparently faults were
lithified or intruded by dikes and thus are more resistant to erosion than
the country rock. Conversely, the observation implies that the country
rock is weaker relative to the faults. Such weak country rock would be
consistent with wall rock composed of breccia [10] that is weakly cemented
by ice near the free faces and is charged with water at some depth.
Another argument supports the idea that the wall rock contained water
and ice. Dark deposits interpreted as volcanic-vent material [5,6] occur
only at elevations lower than 6 km above Martian datum. The highest
deposits are 3 km below the rim of adjacent plateau surfaces. This 5-km
elevation appears to be the maximum height reached by extruding magmas and
can be used to calculate relative densities of magma and wall-rock columns
[14jl5], It appears that the material in the column of combined solid
crust and mantle rock underlying the plateau must have been less dense than
the material in the liquid magma column. Upper crustal rock composed of
loosely consolidated breccia mixed with water or ice might fulfill such a
requirement.
Because the main evidence for water and ice in the wall rock comes
from landslides, their time of emplacement is important. The landslides in
Valles Marineris date from the time of late eruptions on the Tharsis
volcanoes [2] and thus were emplaced after the major activity on Martian
outflow channels, Thereforej the concept of ground saturated by water and
ice in the equatorial region is consistent with Carr's [16] hypothesis that
confined aquifers developed in this region and gave rise to outflow
channels. The concept also agrees with the presence of rampart craters in
the equatorial area.
None of the above observations conclusively demonstrate that water or
ice existed in the wall rock of Valles Marineris, but altogether the
evidence is highly suggestive. Any models addressing the exchange of water
with the atmosphere in the equatorial region of Mars must therefore take
into account that, below a depth of about 1 km, this region was not
entirely desiccated, at least until the time of landslide formation.
References
1)
Fanale F, R, (1976) Icarus, 28, p, 179-202,
2)
Lucchitta B, K. (1979) J, Geophys, Res., 84, p, 8097-8113.
3)
Lucchitta B, K,, Kaufman K« L,, and losllne u, J, (1981) NASA IM
84211, p, 326-328,
4)
Lucchitta B. K. (1981) NASA TM 84211, p. 419-421,
5)
Lucchitta B, K, (1985) Lunar Planet. Sci, XVI, p. 503-504.
6)
Lucchitta B, K, (1986) Lunar Planet. Sci, XVII, p, 496-49/.
7)
Lucchitta B, K, (1984) Workshop on Water on Mars, Nov, 3U-Uec. l
1984, p. 45-47.
8)
Pieri David (1980) Science, 210, p. 895-897.
9)
Scheidegger A. E. (1973) Rock Mechanics, 5, p. 231-236.
10)
Hsii K. J, (1975) Geol , Soc, Am. Bull., 86, p. 129-140,
314
11) Plafker George and Erickson G, E. (1978) Rock slides and avalanches ,
Barry Voightj eda^ Elsevier^ Amsterdam, p. 277-314,
12) Lucchitta B. K. and Ferguson H, M. (1983) Proc. Lunar Planet, Sci.
Conf, 13th , Part 2, 88, Supplement , p. A553-A568,
13) Lucchitta B, K. (1982) Repts. TTifTet. GeoU Prog, -1982: NASA TM
85127 , p, 233-234,
14) Eaton J. P, and Murata K, J, (1960) Science , 132 , p. 925-938.
15) Vogt P, R, (1974) Earth and Planet, Sci. Letts, , 23^, p. 337-348.
16) Carr M. H, (1979) J. Geophys. Res, , 84, p, 2995-3007,
*(Pub1ished in Symposium on Mars: Evolution of Its Climate and Atmosphere,
July 17-19, 1986, p, 59-61,)
315
FORMATION OF THE LAYERED DEPOSITS IN THE VALLES MARINERIS,
MARS
Susan S. Nedell, NASA Ames Research Center, MofFett Field, CA 94035, and Steven W.
Squyres, Cornell University, Ithaca, NY 14853.
Thick sequences of layered deposits are found in the Valles Marineris, Mars [1-3] . They exhibit
fine, nearly horizontal layering, and are present as isolated plateaus of what may once have
been more extensive deposits. Individual sequences of layered deposits are as thick as 5 km.
We have argued previously that the morphology of the deposits is most consistent with origin
in standing bodies of water [3]. The rhythmic nature of the layering, their lateral continuity,
horizontality, great thickness, and stratigraphic relationships with other units in the canyons
all appear most consistent with deposition in a quiet aqueous environment. If standing bodies
of water existed for any significant period of time in the Valles Marineris, they were almost
certainly ice-covered. Here, we examine in detail the conditions necessary for the existence of
ice-covered martian paleolakes, and consider mechanisms for sediment deposition in them.
Ground water has been very important in shaping the martian surface. There is ample evidence
that the martian regolith is highly porous and permeable, and that it at one time contained
large amounts of water [4] . If a large tectonic depression such as the Valles Marineris cut deep
into an aquifer system in the martian regolith, it would be natural for the canyons to become
partially filled with water. As long as the subsurface aquifer system remains charged, water in
the lakes could be replenished readily by seepage from the canyon walls.
It is unlikely that the martian atmosphere was thick enough to have sustained mean annual
temperatures above freezing after the earliest epoch of martian history. Any standing water
bodies would be expected to have a perennial ice cover. Perennially ice-covered lakes presently
exist in the Dry Valleys of Antarctica. There the mean annual temperature is also well below
freezing, and the ice on the lakes has reached an equilibrium thickness [5]. As ice is lost from
the upper surface of the lake by ablation, new ice forms at the lower surface, releasing latent
heat as it does. This heat is the dominant term in the energy balance equation that gives the
equilibrium ice thickness. Water in the lake is replenished in Antarctica by surface flow; on
Mars this could be accomplished by ground water seepage. The rate of ablation exerts a strong
influence on the equilibrium ice thickness, and on Mars it is poorly known. Based on reasonable
estimates of ablation rates, the equilibrium thickness of ice on martian lakes under the present
climate might be 65 to 650 m [5]. The depth of possible lakes is very poorly constrained, and
could have ranged from very shallow depths to more than 5 km deep.
There are three ways that sediment could enter an ice-covered lake: down through the ice
cover, up from the lake bottom, or in from the lake margins. We now consider each of these
possibilities in turn.
Four processes that could have transported sediment downward through an ice cover are con-
sidered: (a) solar energy warmed individual particles, allowing them to melt through the ice;
(b) sediment worked its way downward through vertical melt channels; (c) a layer of sediment
deposited on the ice was thick enough to cause the ice layer to founder, dumping the sediment
into the lake; and (d) a layer of sediment deposited on the ice cover led to a Rayleigh-Taylor
instability, and sediment diapirs penetrated downward through the ice layer.
Simple energy calculations show that solar warming is inadequate to melt moderate-sized grains
downward through the 5-meter ice cover of typical Antarctic lakes [6]. It is therefore expected
to be wholly inadequate on Mars, where the surface temperature and solar flux are still lower,
and the equilibrium ice thickness may be one or two orders of magnitude greater.
316
Migration through vertical melt channels appears to be the primary sedimentation mechanism
in some Antarctic lakes [7] . Melt channels form when the ice thickness is less than about 3 m. In
order for this process to have operated on Mars, surface temperatures and ablation rates must
have been high enough to thin the ice to only a few meters, and liquid must have been stable
at the surface for periods during the summer. Neither condition is likely to have been met, as
there is no evidence that the martian climate at the time of layer deposition was substantially
warmer than it is at the present.
Foundering of the ice cover could have occurred if enough debris was loaded onto the ice surface
so that the overall density of the ice-sediment layer became greater than that of the liquid
below. As sediment accumulated on the ice surface, the ice layer would thicken continuously,
since freezing at the lower surface of the ice would no longer be balanced by ablation from the
upper surface. We have calculated, for a given thickness of the sediment layer to be dumped
into the lake, how rapidly sediment must be piled onto the ice surface for foundering to occur.
We take ice surface temperatures of 210 to 240 K, bulk sediment densities of 1.5 to 2.5 g cm~^,
and sediment layer thicknesses of 75 to 150 m (the thicknesses of a light layer and a light/dark
couplet, respectively, measured in Candor Chasma). Sedimentation rates of 0.4 to 15 mm yr~^
are required, and will lead to foundering at ice thicknesses ranging from 0.5 to 2.8 km. A
weakness of the foundering hypothesis is that the ice would not undergo substantial melting
during the foundering event, and could subsequently reform as a continuous cover and continue
to thicken. However, these calculations neglect the effect of a geothermal heat flow. If the
geothermal heat flow had an Earth-like value, the equilibrium thickness of even a sediment-
covered ice layer naight be no more than ~ 2 km. Foundering could then occur repeatedly,
taking place each time the sediment thickness exceeded the critical value. For an equilibrium
ice thickness of 2 km, continuous sedimentation and repetitive foundering with a bulk sediment
density of 2.0 g cm~^ would produce a sequence of layers each 140 m thick.
Even if foundering did not take place, it is likely that a thick layer of sediment that accumulated
on the ice surface would penetrate the ice via a Rayleigh- Taylor instability. The configuration
of dense sediment over less dense ice would favor diapiric upwelling of the ice and sinking of
the sediment. We consider the flow to be dominated by the rheology of the ice, and take an
upper sediment layer 75 to 150 m thick. For an ice temperature near freezing, the instability
will grow in tens of years. For a temperature of 210 K, the growth time is of the order of 10*
yr. Therefore, if a sediment layer thick enough to form one of the observed layers accumulated
on an ice cover, it would probably penetrate it rapidly.
The limiting factor for sediment deposition by foundering or Rayleigh-Taylor instability is the
ability to accumulate substantial amounts of sedimient on the ice surface. Global dust storms
could conceivably be the source for the sediment, but there are significant problems with this
hypothesis. The present martian climate produces net deposition of dust at the poles, and
this process would have to be somehow reversed. Furthermore, sediment would have also
presumably accumulated on the surrounding uplands as well as in the Valles Marineris. None
is presently observed there. This difliculty could be overcome if there were repeated periods of
deposition and erosion near the equator. Sediment built up on the uplands would be swept away
during erosional episodes, while debris deposited on the ice would be trapped by foundering or
Rayleigh-Taylor instability, and preserved. Without a clearly plausible mechanism for massive,
repeated sedimentation at low latitudes, however, origin of the deposits by downward migration
through an ice cover remains speculative at best.
Volcanic material that originated beneath the lake might also be a source for the layered de-
posits. There are several arguments against formation of the deposits by subaerial vulcanism.
One would expect that accumulations of ash-fall debris would be widespread, yet there are no
317
layered deposits on the uplands surrounding the Valles Marineris. The nature of the layering
also does not support an ash-flow origin. Typical large terrestrial ash flows form aprons of
individual flows that taper away from central vents. Smaller and more abundant flows would
produce irregular layering. In addition to there being no unambiguous volcanic calderas, the
layered deposits are characterized by fairly uniform layer thicknesses that extend laterally for
at least many tens of kilometers. These arguments would be largely overcome if the vulcanism
were subaqueous. Volcanic constructs miay have been destroyed by slumping of material off
cones as they were forming, or masked by floating pumice that eventually became waterlogged
and sank to the lake bottom. Volcanic eruptions in water also would distribute effusive mate-
rial much more easily. Even at fairly great depths, it is likely that explosive vulcanism could
occur on Mars. On Earth, eruptions change from effusive to explosive activity, due to magma
vessiculation, at water depths of 300 m for basaltic magma [8] , and 500 m for silicic magma [9] .
The corresponding water depths for explosive eruptions on Mars are about 800 m and 1300 m,
respectively. Although there is no direct evidence for it, the process of subaqueous explosive
vulcanism is an attractive mechanism for explaining some aspects of the layered deposits.
Finally, the nearby canyon walls are an obvious source of sediment for the layered deposits. It
is likely that the Valles Marineris formed as tectonic grabens, some of which were substantially
enlarged by removal of interstitial ground ice and collapse of the canyon walls. In a lacustrine
environment, sediment would have been transported from the canyon walls into the deeper por-
tions of the canyons by gravity flows, and deposited in nearly horizontal layers. This mechanism
presents some geometric complications, and may not be able to account for all of the deposits.
The material from the collapsing canyon wall is sufficient to fill the depressions that formed
only partially, yet the present deposits rise nearly to the level of the canyon rims in places. This
problem could be alleviated if plateaus of layered deposits had cores of undisturbed canyon wall
material, or if material were also added by volcanic eruptions or sediment transport downward
through the ice.
We conclude that there are several geologically feasible mechanisms that could have led to
formation of thick deposits in ice-covered paleolakes in the Valles Marineris. Present data are
insufficient to choose conclusively among the various possibilities. Several types of data from
the Mars Observer mission will be useful in further characterizing the deposits and clarifying
the process of their origin. The deposits should be considered important targets for a future
Mars sample return mission.
References: [l] McCauley, J.F., USGS Misc. Inv. Ser. Map 1-897 (1978); [2] Lucchitta, B.K., NASA
Tech. Memo. 85127, 233 (1982); [3] Nedell, S.S., and Squyres, in Workshop on Water on Mars, 56 (1984);
[4] Carr, M.H., J. Geophys. Res. 84, 2995 (1979); [5] McKay, C.P., et al, Nature 813, 561 (1985); 6
Simmons, G.M., et al., U.S. Antarc. Jour. (1985); [7] Nedell, S.S., submitted to Sedimentology; [8
Sigurdsson, H., Geol. Assoc. Can., Short Course Notes 2, 294 (1982); [9] Moore, E.G., and Schilling,
J.G., Contrib. Min. Petrol. 41, 105 (1973).
318
QEOMORPHIC EVIDENCE FOR ANCIENT SEAS IN WEST DEUTERONILUS MENSAE, MARS - 1:
REGIONAL 6EOMORPHOL(X3Y
by
Timothy J. Parker. Dale M. Schneeberger, David C. Pieri, and R. Stephen Saunders
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA
Oeuteronilus Mensae lies at the western end of a swath of fretted terrain greater than 500km wide and
3000km long comprising the boundary between the northern lowland plains of Mars and the topographically higher
and more heavily cratered southern highland terrain from 30° to 50' latitude, 280* to 350* longitude. We will
concentrate on the fretted terrain morphology within a specific region in west Deuteronilus Mensae (44* to 50'
latitude, 342* to 347' longitude, figure 1).
The fretted terrain in west Deuteronilus flensae consists of extensive cratered upland "penninsulas" or
isolated plateaus cut by long, finger-like canyons typically 10 to to 20km wide and upwards of 300km long. The
longest of these canyons trend roughly north-south to north-northeast in much of the region depicted in Tigure 1 .
which may reflect some local structural and/or topographic control. Along the east side of figure 1, and
throughout much of Deuteronilus Mensae. the lowland dissection of the cratered upland also exhibits some circular
and arcuate trends that might indicate preferential degradation of buried large impact structures (1). This
particular example of marlian fretted terrain presents a relationship to the lowland/upland boundary which is
unique in that the gross fretted terrain morphology, rather than defining the lowland/upland boundary as is the
case for most martian fretted terrains, is in fact present on both sides of the boundary (though it is highly
degraded on the north side).
At least three geomorphic zones roughly parallel to the lowland/upland boundary, suggestive of increasing
modification northward, can be recogni£ed on the fretted terrain of the region. These zones appear to be
equivalent to what Weiss et al. (2) interpreted as stratigraphic units. For convenience of description we have
identified these zones, from southern highlands to northern plains, as "A°, "B", and "C".
The southern-most zone (zone 'A") consists of sharply defined fretted terrain. The highland plateau surfaces
generally appear brighter and more varied in albedo and texture than do those of the middle zone immediately to
the north (Zone "B"). They are separated from the canyon floors by escarpments ranging from 262m (±32m) near
the plateau surface contact with zone"B", to greater than 650m {±32m) toward Uie south edge of figure 1 . (All
shadow measurements for the canyon walls were measured from Viking O^biter image 673B43. which covers this
same region at a resolution of 204 m/pixel with a sun elevation of 9'). The canyon floors of zone "A" are
comprised of smooth plains partially to completely buried by relatively bright debris aprons from the canyon
walls. These debris aprons occur as sharply defined, gently sloping surfaces that are either concentric to isolated
plateau outliers or parallel to canyon walls and are typically less than lOkm wide. They occur at the bases of all
fret escarpments within zone "A" except those within about 50km of the contact with zone "B" on the plateau
surface. The contact between zones "A" and "B" is a well defined irregular line running east-west across the
cratered upland surface and plateau outliers. This contact can be readily traced across the highland surface for
several hundred kilometers, both to the southwest and east of the area shown in figure 1 . (^ the lowland surface
it is less easily recognized, particularly to the east, where the plateau outliers are more scattered.
The middle zone (zone "B") consists of well, defined fretted terrain in which the plateau surfaces appear
smoother, wiUi a somewhat darker and much less varied albedo surface than those of zone "A". The canyon walls
in zone "B" rang® in height from less than 54m (±32m) just north of the degraded 10km crater at D in figure I , b
as high as 163m (±32m) north of the plateau surface contact between zones "A" and "B". North of the crater at D.
the canyon wall slopes ^proach that of the sun elevation, so heights cannot be determined reliably by shadow
measuremanb. Ths canyon floors of zone "B" are comprised of smooth plains similar to those of zone "A", but
lack the prominent debris aprorts. Instead, much of the canyon floors adjacent to the escarpments either lack
debris aprons entirely at this scale or, at best, exhibit poorly defined or subdued debris aprons expressed as
narrow features (a few km or less wide) intermediate in slope between those of the canyon wall and floor.
The northern-most zone (zone "C") consists of rounded or "softened" fretted terrain. The plateau and canyon
surfaces consist of light and dark "striped" terrain within about 50km of the contact with zone "B". and mottled
terrain beyond about 50km from the contact with zone "B". Slope inflections at plateau/canyon margins within
zone "C" are very subdued. On the plateau and upland surface, the contact between zones "B" and "C" is expressed
as a moderately well defined line separating the smooth, relatively uniform albedo surface of zone "B" from the
striped or mottled surface of zona "C". Within the fret canyons, the contact between zones "B" and "C" is
somewhat sharper than it is on the plateau surface and is expressed as smooth arcs or lobes with their concave
sides facing zone "C" (at E. figure 1). The plateau/lowland morphology typical of the fretted terrain, though
319
apparently highly subdued, is visible in this region as far north as 51' latitude, more than 200km north of the
contact between zones "B" and "C" - approximately coincident with the usual position given for the boundary
between the northern plains and the southern highlands (3,4).
Weiss et al. (2) interpreted the zones as surface exposures of successively lower stratiqraphic units. In their
model, these units are nearly horizontal layers intersecting the northward-dipping plateau surface. However, at
least two problems with a stratiqraphic interpretation for the zoning become apparent in the 200m/pixel images of
figure 1 Such problems become even more apparent upon examination of very high resolution images (see
footprints for these images on map. figure 1 ) of the region (5): (A) The northern limit of the occurrence of
prominent debris aprons associated with zone "A" exhibits an apparently topograpically conformal offset to the
south with respect to the plateau surface contact between zones "A" and "B". If the debris aprons are associated
with the wasting of an upper "A" stratigraphic unit, they should occur as far north as the plateau surface contact
between zones "A" and "B". Also, ihs gradual decrease in height of the canyon walls to the north might expectedly
produce progressively narrower debris aprons to the north. Further, if they are comprised of material from all
three units, one might expect them to be found associated with escarpments in zones "B" and "C" as well. (B) The
smooth, lobate contact between zones "B" and "C" (at F, figure 1) embays an old, degraded crater about I5km in
diameter. A stratigraphic contact would have been disrupted by the formation of the crater.
Another, though provocative mechanism by which topographically conformal zone contacts might be produced
on the fretted terrain is by successively lower levels of standing water associated wiUi episodes of catastrophic
outflow channel development elsewhere along the lowland/upland boundary. If the zone boundaries represent old
shorelines, the above problems can be addressed: (A) The absence of prominent debris aprons in zones "B" and
"C" could be due to reworking or complete removal of the debris by rising and falling water levels. The southward
offset of the contacts between zones "A" and "B" and between zones "B" and "C" could be due to successive
episodes of embayment of the canyons. (B) Embayment of the 1 5km degraded crater by the contact between zones
"B" and "C" easily fits the expected behavior of a shoreline.
In addition to providing reasonable explanations for the above problems, successive levels of standing water
within the northern lowlands might be useful in addressing some other fretted terrain problems. (A) Removal of
debris (likely to be comprised of a wide range of grain sizes) wasted and/or sapped from the canyon walls in a
near shore lacustrine environment would be efficient arid could produce the smooth canyon floor surface without
affecting the plateau surface (as might be expected of eolian processes), it also provides a very effective way of
maintaining the steep cliffs by focusing erosion (through wave action) at the bases of the cliffs and, at the same
time, preventing the accumulation of talus deposits.
(1) Sharp, P.. P., 197.3. Mars: Fretted and chaotic terrains: Journ. Geophys. Res., Vol. 78, p. 4073-4083.
(2) Weiss, D.. Fagan. J. J., Steiner, J., and Franke. 0. L.. 1991, Preliminary obsevations of the detailed
stratigraphy across the highland-lowlands boundary: Reports of the Planetary Geology Program - 1981,
NASA Tech. Memo. 842! 1, p. 422-425.
(3) Lucchita, B. K.. 1978. Geologic map of the Ismenius Lacus (Quadrangle of Mars: USGS Map 1-1065. Atlas of
Mars, 1:5.000.000 Geologic Series.
(4) Scott, D. H., and Carr, M. H., 1978, Geologic map of Mars: USGS Map 1-1083, Atlas of Mars, 1:25.000.000
Geologic Series.
(5) Parker, T. J., Schneaberger. D. M., Pieri, D. C and Saunders, R. S., Geomorphic evidence for ancient seas in
west Deuteronilus Mensae. Mars - II: From very high resolution Viking Obiter images: This volume.
Figure I: Geomorphic Map of West Deuteronilus Mensae, Mars. Viking Orbiter Images: 675634,59.
320
I i* lifs {ill * "ill I I
jsij5i5S| i,;;jts.5 jjjsti I If! fl'i
Hililill iUsiil i$£S = 5 - til f?s
- i '»--lFI- i-l-i- s"sl
li SI His
if! Iliitf " fli llj
fill
III
..-, \.
C/^»».?-p^'.-fr, '-1,
\'* '-•'.€:
321
0EOMORPH1C EVIDENCE FOR ANCIENT SEAS IN WEST DEUTERONILUS MENSAE, MARS - li;
FROM VERY HIGH RESOLUTION VIKING ORBITER IMAGES
by
Timothy J. Parker. Dale M. Schneeberger. David C. Pieri. and R. Stephen Saunders
Jst Propulsion Laboratory, California Inatibut* of Technology, Pasadena, CA
Very high resolution (lOm/pixel or better) Viking Orbiler images of the martian surface, though rare, make it
possible to examine specific areas at image scales approaching those of high altitude terrestrial aerial photographs.
Of the approKimately 1300 total Viking images in this range, probably less than 500 of them are clear enough to be
very useful for studies of the martian surface. Twenty three clear very high resolution images (458B41-51,
61-72; at 9m/pixel) lie within west Deuteronilus Mensae (see 1 figure 1, for location). This discussion will deal
with the northernmost images (458B61-72) which constitute an almost unbroken mosaic of the west wall of a long
finger-like fret canyon (figure 1 ).
In the vBvy high resolution images, morphological details on the plateau surface within zone "B" (1), not
detectable at low resolution, make it possible to divide the zone into two distinct subzones separated by an
east-west escarpment (at D. figure 1). South of this escarpment, the plateau surface is characterized by its
smooth texture at small scales and relatively uniform albedo. The plateau surface north of this escarpment has a
more mottled appearance. The most notable characteristic of this subzone is the presence of well defined polygons
several tens of meters to as much as 200m to 300m across (2). These are best expressed on the plateau surface
from about 2km north of the escarpment at D to within 20km of the top of figure 1 . Individual polygons are
separated by well-defined cracks that are typically a few tens of meters or less wide and appear filled with
relatively bright material. The canyon floor south of the contact between zones "B" and "C" is characterized by
the presence of polygons similar to but somewhat larger than those on the plateau surface.
The canyon floor north of the contact between zones "B" and "C° consists of two principal morphologic styles.
Much of it consists of a relatively uniform albedo, smooth plains surface, broken by a system of narrow, sharp
wrinkle ridges. The remainder of the canyon floor consists of a system of bright, low relief mounds. Their
contacts with the surrounding plains are usually quite sharp. Individual mounds are typically ovate to elongate
with a rounded, irregular outline. The ovate mounds are up to 300m across. The elongate mounds may be 300m
wide by as much as 3km long and occur as disconnected, subparallel arcs concave northward. It is this system of
subparallel arcs which gives the southernmost portion of zone "C" its "striped" appearance in lower resolution
images ( 1 , figure 1 ).
0>e of the more inbresting aspects of this set of images is the presence of at least three, and possibly four
parallel, and apparently topographically conformal benches on the canyon wall. These benches become
progressively less sharply defined from lowest to highest (numbered 1 through 4), and are traceable for many
tens of kilometers along the canyon wall. Only the lowest of these, bench 1, can be seen in the lower resolution
images as the contact between zones "B" and X". Each bench occurs progressively higher up the wall when traced
from south to north. Benches 2, 3, and 4 each in turn disappear at the top of the escarpment. Bench 2 intersects
the top of the escarpment in the vicinity of the fresh 800m crater near the top of figure ! . Bench 3 intersects the
cliff top just south of the 1 0km degraded crater in figure 1 . Bench 4. less readily traceable than the other three
benches, appears to coincide with the base of the east-^west escarpment on the plateau surface. Bench 1. though it
cannot be seen inlsrsacling th® cliff top, is the only one of the four that intersects the canyon floor in the very
high resolution imaps.
The contact feiiwten zones "B" and "C" on the canyon floor and along part of the canyon wall is well
represented in the vary high resolution images. On the canyon floor, it consists of a parallel system of alternating
low ridgss and swaiss sbmit !km wide (E. figure 1) crossing the canyon floor. These describe smooth parallel
arcs, concave northward. At its west end, the southernmost of the arcs, a low positive relief feature, grades into
bench 1 at the base of the canyon wail. From here northward, it occurs at successively higher positions on the
canyon wall until, at the top of figure 1. it appears to be about midway up the wall.
Shading across the benches suggests that the steepest slope elements occur immediately above the benches,
with the profile between benches being convex upward. This is opposite what might be expected of most
stratigraphic benches. The most common profile for a cliff comprised of layers of varying resistance to erosion is
concave upward for the slope profile between benches, in such an example, the steepest slope element lies
immediately below the bench (since the more resistant layers tend to be cliff formers). Sapping interfaces within
a cliff might produce profiles similar to those in west Deuteronilus flansae. In this case, sapping above a
permeability boundary results in more rapid erosion immediately above the boundary then below. Sapping,
however, does not adequately explain the change in the contact between zones "B" and "C" from a bench to a series
322
of arcuate ridges and swsles across ths canyon floor.
A series of strandlinas, each lower and younger than the previous one could explain all the above aspects
Wave acti(Mi against g slopsd sirfac® (in this case the canyon wall) could produce the observed profiles. Wave
energy is focused m Oj® si(^ at the shorezone. producing a cliff on the slope immediately above. If successively
lower "high steuls" mrs ssparalsd by enough time, the lowest, youngest strandline should be the best preserved
and the highest the least preserved. The series of ridges and swales at the contact between zones "B" and "C
similar in scale and plan to terrestrial beach ridges on a very gentle slope.
are
REFERENCES:
(1) Parker, T. J., Schneeberger, D. M., Pieri. D. C, and Saunders, R. S., Geomorphic evidence for ancient seas in
westDeuteronilusMensae. Mars-I: Regional Geomorphology: This volume.
(2) Lucchita, B. K., 1984, Small-scale polygons on Mars: Reports of the Planetary Geology Program - 1983
NASA TM 86245, p. 205-207.
Figure I: Very high resolution photomosaic of part of west Deuteronilus Mensae, Mars. Viking Orbiter images:
458861-72. Regional slope is downward to the north. Brief unit descriptions: (Pp), Polygonally patterned
plains; (Ps), Smooth Plains; (Psp), "Splotchy" smooth plains; (S), Bright, low mounds (splotches); (B).
Blocky material; (Cs). Canyon wall and/or cliff slopes; (Ds). "Debris" slopes; (Ce). Crater ejecta blankets-
(H), Hilly material; (C), Craters - C4 freshest, CI most degraded.
323
Zone C
324
CHAPTER 10
VOLCANIC PROCESSES AND LANDFORMS
EXPLOSIVE VOLCANIC DEPOSITS ON MARSj PRELIMINARY
INVESTIGATIONS
DA. Crown^ LA. Leshm, and R. Greeley, Department of Geology, Arizona State
University, Tempe, Arizona 85287
The existence of explosive volcanism and the identification of related eraptive
products on Mars have been the subject of much controversy. Most research regarding
martian volcanism has centered around basaltic volcanism. Terrestrial analog studies,
photogeologic investigations, Viking Lander chemical analyses, and remote sensing
information all indicate the presence of basaltic volcanism on both regional and local scales
[1]. Small-scale explosive volcanism is indicated by possible cinder cones [2-4] and
"pseudo craters" resembling those in Iceland [5]. A possible ash-fall deposit associated
with Hecates Tholus has been proposed [6], and volcanic density currents have been
suggested to explain features surrounding volcanoes in the Tharsis region [7]. The
possibility of explosive volcanism on a larger scale was implied by Malin's [8] comparison
of the Elysium region of Mars and the Tibesti region on Earth. Other investigators have
considered the existence of rhyolitic lava flows [9] and silicic domes [10], but the
occurrence of silicic volcanism, which is commonly explosive in nature on tike Earth, has
received little support [11]. Large-scale explosive volcanism was suggested to account for
the basal scarp [12] and aureole deposits [13] of Olympus Mons, as well as deposits
covering > 10^ knfi in the Amazonis, Memnonia, and Aeolis regions [14,15]. The
deposits in Amazonis Planitia were proposed to be pyroclastics on the basis of similarities
(such as complementary joint sets and thick flow sheets) to ignimbrites in the Pancake
Range of Central Nevada [15], but the morphology observed on Mars may not be definitive
of ignimbrites, and the deposits have also been suggested to be of a non-volcanic polar
origin [16] and mantling deposits of an aeolian origin [17]. The formation of the martian
highland paterae, which are distinguished by low, broad forms, radial channels, lobate
ridges, and central caldera complexes [18], has been attributed to explosive volcanism.
Originally, Potter [19] suggested that the highland paterae were formed by eruptions of
extremely fluid basaltic lava, but Pike [20] noted the morphometric correlation of the
highland paterae with large ash sheets on Earth. Greeley and Spudis [1] proposed that the
highland paterae were formed by phreatomagmatic eruptions in an early period of martian
history.
Two investigations have been undertaken to examine possible large-scale explosive
volcanic deposits on Mars. The first includes an analysis of Viking Infrared Thermal
Mapper (IRTM) data covering the vast deposits in the Amazonis, Memnonia, and Aeolis
regions. These "postulated ignimbrites" have been mapped by Scott and Tanaka [15], and
at least five high-resolution nighttime IRTM data tracks cross the deposits. Preliminary
analysis of the data covering Amazonis Planitia show that local features (such as yardangs)
have anomalous thermal inertias but the "ignimbrites" as a whole do not consistently have
significantiy different thermal inertias from their surroundings. However, this does not
necessarily discount the existence of ignimbrites, as local features and apparent mantUng by
a fine layer of dust can dominate thermal inertias [21], Further investigation including
examination of lower resolution data is needed to provide a complete characterization of
these deposits using IRTM data and to ascertain whether they possess a discemable thermal
signature.
Preliminary photogeologic and IRTM studies of the large and small highland
paterae [22] have also begun. Two high-resolution nighttime ttacks trending NW-SE cross
ApoUinaris Patera, which is located adjacent to the lowland-highland scarp (9°S, 186°W).
In both the northern and southern regions of ApoUinaris Patera thermal inertias are
327
significantly greater than those observed in the surrounding areas. The thermal signature of
ApoUinaris Patera is more prominent on its southern flanks; within the southernmost of the
two tracks thermal inertias increase from ~ 3.0 - 3.5 in the region surrounding ApoUinaris
Patera to a high of greater than 6.0 (x 10"^ cal cm-2 sec-l''2/K-l) on the volcano and within
the northern track from ~ 1.9 to a high of 3.4. Lower thermal inertias on the northern
flanks of ApoUinaris Patera are consistent with the regional trend reflecting the transition
from the cratered uplands to the smooth plains [21] and also with observations of less
morphologic detaU on the northern flanks suggesting more thoroughly mantled surfaces.
Continued analysis of IRTM data covering Uie large and small highland paterae will
determine if the paterae possess distinctive thermal signatures.
The purpose of IRTM studies of postulated martian explosive volcanic deposits is
to determine the physical properties of the proposed ignimbrites. If volcanic deposits are
exposed at the surface, high thermal inertias, as are observed for ApoUinaris Patera, should
be present. However, the thermal signature of the paterae, for instance, may not be the
result of a primary volcanic surface. If a distinctive thermal signature is observed, it may
be only an indirect reflection of the volcanic deposits. Nevertfieless, this would allow a
characterization (by albedo, thermal inertia, and rock abundance [23]) of the material
controlling the thermal properties of the paterae and should provide important information
for the consideration of explosive volcanism on Mars.
Although many investigators have considered explosive volcanism on Mars,
currently much of the evidence for extensive deposits is tenuous, and the occurrence of
large-scale explosive volcanism remains controversial. The present studies are intended to
address the question in more detail with the aid of new and previously unused sources of
information. The implications of explosive volcanism are critical to understanding the
evolution of the martian surface. In combination with the examination of possible martian
explosive volcanic deposits, this investigation wUl also include a remote sensing analysis
and a field examination of a selected terrestrial analog. Due to their morphologic
similarities, ignimbrites in the Central Andes are presently the best candidates as analogs
for the martian highland paterae.
R^erences
I] Greeley, R., and Spudis, P.D., 1981, Rev. Geophys. Space Phys., 19, 13-41.
2] West, M., 1974, Icarus, 21, 1-11.
3] Mouginis-Mark, P.J., 1981, Proc. Lunar Planet. Sci. Conf., 12th, 1431-1447.
4] Lucchitta, B.K., 1985, Lunar Planet. Sci. Conf., XVI, 503-504.
5] Frey, H., and Jarosewich, M., 1982, /. Geophys. Res., 87, 9867-9879.
6] Mouginis-Mark, P.J., WUson, L., and Head, J.W., 1982, /. Geophys. Res.,87,
9890-9904.
7] Reimers, C.E., and Komar, P.D., 1979, Icarus, 39, 88-100.
8] Malin, M.C., 1977, Geol. Soc. Am. Bull., 88, 908-919.
9] Fink. L, 1980, Geology, 8, 250-254.
10] Plescia, J.B., 1981, Icarus, 45, 586-601.
II] Francis, P.W., and Wood, C.A., 1982, /. Geophys. Res., 87, 9881-9889.
12] King, IS., and Riehle, J.R., 1974, Icarus, 23, 300-317.
13] Morris, E.G., 1980, NASA TM-82385, 252-254.
14] Scott, D.H., and Tanaka, K., 1980, NASA TM-82385, 255-257.
15] Scott, D.H., and Tanaka, K., 1982, /. Geophys. Res., 87, 1179-1190.
328
[16] Schultz, P.H., and Lutz-Garihan, A.B., 1981, Lunar Planet. Sci. Conf., XII, 946-
948.
[17] Greeley, R., Williams, S.H., White, B.R., Pollack, J.B., and Marshall, J.R., 1985,
in Models in Geomorphology, ed. MJ. Woldenberg, Allen & Unwin, Boston, 373-
422.
[18] Plescia, IB., and Saunders, R.S., 1979, Proc. Lunar Planet. Sci. Conf., 10th,
2841-2859.
[19] Potter, D., 1976, U.S. Geol. Survey Misc. Geol. Inv. Map 1-941.
[20] Pike, R.J., 1978, Proc. Lunar Planet. Sci. Conf., 9th, 3239-3273.
[21] Zimbelman, J.R., and Leshin, L.A., 1986, submitted to Proc. Lunar Planet. Sci.
Conf, 17th.
[22] Albin, E.F., 1986, Master's Thesis (unpublished), Arizona State University, 84 pp.
[23] Christensen, P.R., 1982, /. Geophys. Res., 87, 9985-9998.
329
PSEUDOCRATERS AS INDICATORS OF GROUND ICE ON MARS
Herbert Frey, Geophysics Branch, Goddard Space Flight Center^ Greenbelt, Mn
20771
In the search for reliable indicators of the past location of surface
or near surface volatiles on Mars, pseudocraters (if they exist) would be
of direct but limited use. We have previously suggested that the thousands
of small (subkilometer) pitted cones which dot portions of the plains-
forming units in northern Mars may be volcano-ice analogs of Icelandic
pseudocraters (ls2)s which on Earth form where lava flows over water or
water-saturated ground (3), The steam explosion caused by this interaction
is only marginally less efficient if (as is likely on Mars) ice is the
volatile(4). Positive identification of martian pseudocraters would there-
fore be strong indication of past occurrence of ice at or near the surface
of Mars.
The basis for suggesting that the small cones on Mars are pseudocraters
includes: (a) small size^ (b) abundant but patchy distribution on what
appear to be volcanic plains^ (c) presence of other features suggestive of
surface or subsurface ice. (d) morphological similarities to Icelandic
pseduocraters, and (e) the similarity in distribution of crater/cone diam-
eter ratios to Icelandic pseudocraters(2). This last morphometric parameter
may be the most important^ since other possible small terrestrial volcanic
analogs have very different crater/cone diameter ratio distributions(2).
In a survey of the available high resolution Viking Orbiter imagery, abun-
dant fields of possible pseudocraters were found in SE Acidalia Planitia,
S Utopia-Elysium, W Isidis Planitia and, perhaps, near Hellas (2,5,6),
HoweveTj only a small fraction of the plains-forming units imaged at high
resolution (range < 2000 km) were found to contain the small cones: of
some 12,200 images searched we found subkilometer cones with central pits on
less than 350 (<3%)» This low discovery rate, combined with the limited high
resoluiton imagery, restrict martian pseudocraters as global indicators of
surface or subsurface ice.
There are only minor morphological differences between the subkilometer
cones found in Acidalia and those found in Utopia-Elysium-Isidis; more strik-
ing is the variation in background terrain on which the cones are found. In
Acidalia the cones are found on smooth plains of both uniform and mottled
appearance but also on widespread fractured and subdued fractured plains
(2,7), The fractured plains are of interest because they may themselves be
indicators of ground ice(8). All the plains-forming units on which cones
have been found also contain rampart craters, but there are regional differ-
ences in the size distributions of these craters which may be interesting
(see below)a In Acidalia there are general trends in the dimensional and
distributional characteristics of the cones which seem to depend on back-
ground terrain: the younger plains (which are smooth, not fractured) have
a higher density of cones, but the average cone diameter is lower (<700 m)
than on the fractured plains (>700 m). Cones on subdued fractured plains
(9), which are under-populated in impact craters at the smallest diameters
and have soft, incomplete or interrupted fracture patterns, have the largest
mean diameters observed and also the highest percentage of widest central
craters. Preferential obscuration of small cones and decrease in the
observed cone base diameter due to blanketing by dust seems likely (7).
330
This demonstrates another problem with using pseudocraters as indicators of
ground ice on Mars: because of their small size they are easily removed or
obscured by a variety of erosional processes,
Pseudocraters do have one important contribution to make to the study
of the distribution of volatiles in the martian crust. Because of the way
they forms they indicate the presence of only surface (or very near surface)
ice over which relatively thin lava flows have been emplaced. If the lava
flow is too thicks cones will not form as the work required to lift the
overlying molten rock becomes greater than that available from the explosion,
Likewise, if the ice is buried too deeply beneath an insulating layer, the
heat from the lava flow may dissipate before sufficient volatilization of
the buried ice occurs(4). Therefore the size of pseudocraters and their
spatial density depends on a combination of lava flow thickness and temper-
ature, depth to the ice layer and fraction of ice in mixed layers (soil and
ice). Dense concentrations are favored by relatively thin flows over
abundant ice close to the surface; more wide-spread groupings may indicate
variations in flow thickness and/or the depth to the (top of the) ice
layer.
It would be interesting to compare the spatial density of possible
martian pseudocraters with the size frequency distribution of rampart
craters (10,11,12,13). Not only could such comparisons help to define the
thickness of the ice layer(14), but comparison of rampart and non-rampart
crater populations with a varying density of pseudocraters could place
temporal constraints on the longevity of the ice-rich layer. For example:
in Utopia Planitia where widely scattered small cones occur, rampart cra-
ters are rare for diameters < 3 km, but for D > 5 km almost all craters
have this structure. Perhaps the small craters formed largely after the
near-surface ice vanished. By contrast the rampart craters found in the
fractured plains in Acidalia occur at very small diameters in the regions
where small cones exist.
Despite the sampling problem, the latitudinal distribution of martian
pseudocraters and its comparison with other ice-related features is of
interest. In Acidalia Planitia most of the small cones lie at latitudes
greater than 38°N, ranging up to 50°N, We find no convincing evidence for
such features below ~35°N; however, there is very little high resolution
imagery available. No small pitted ones are found in Chryse between 20
and 30°N, even though good imagery does exist (15,2), By contrast small
cones are found as low as 10°N latitude in Isidis, These are unusual in
their spatial distribution, however, being very densely grouped and often
occuring in long chains(2). If these are also pseudocraters, then at the
time they formed surface or near surface ice must have existed at this low
latitude in eastern Mars.
References :
1, Frey, H, et al,, J. Geophys. Res. 84 , 8075 (1979),
2, Frey H. and M. Jarosewich, J. Geophys. Res. 87 , 9867 (1982),
3, Thorarinsson, S., Bull . Volcanol . Ser, 2 , 14 , 3 (1953).
4, Rocha, D., unpublished manuscript (1979),
5, Frey, H, and M, Jarosewich, Lunar Planet, Sci . XII , 297 (1981),
6, Frey, H, et al.. Lunar Planet. Sci. XII , 300 (1981)",
7, Frey, H. and A.M. Semeniuk, Lunar Planet. Sci. XV , 272 (1984).
331
8. Carr, M.H. and G.G, Schaber, J, Geophys, Res. 82 , 4039 (1977)
9, Guest, J.E. et al . , J. Geophys. Res. 82 , 4111 (1977).
10. Johansen, L.A., NASA TM 80339 , 123 (1979).
11. Allen, C.C, Icarus 39 , 111 (1979).
12. Carr, M.H. et a1., J. Geophys. Res. 82 , 4055 (1977).
13. Mouginis-Mark, P.J., J. Geophys. Res. 84 , 8011 (1979).
14. Barlow, N.G., B.A.A.S7T7 , 735 (1985).
15. Greeley, R. et a1 . , J. Geophys. Res. 82 , 4093 (1977).
332
ERUPTI¥E HISTORY OF TOE ELYSIUM ¥OLCAHIC PROVINCE OF HARS
Kenneth L. Tanaka and David H, Scott, U.S. Geological Survey^ Flagstaff, AZ
New geologic mapping of the Elysium volcanic province at 1:2,000,000
scale and crater counts provide a basis for describing its overall eruptive
history. (We counted craters larger than 2 km in diameter to achieve optimal
consistency; accurate counting of smaller craters is adversely affected by
apparent variations in density due to secondaries, volcanic pits, image
quality, and erosion.) The following stages are described in order of their
relative age; they are also distinguished by eruption style and location.
Stage 1: Central volcanism at Hecates and Albor Tholi
These volcanoes are embayed by lava flows from Elysium Mons. This
relation, as well as crater densities (Table 1), supports an Upper Hesperian
surface age [1]. We have no evidence to determine when these volcanoes became
active. Their steep slopes indicate that they are composed of once-viscous
lava or interbedded lava and pyroclastic material [2,3]; no lava-flow scarps
were identified.
Stage 2: Shield and complex volcanism at Elysium Mons and Elysium Fossae
The vast majority of exposed lava flows in the region were extruded from
Elysium Mons and nearby fissures. The flows overlie Lower Hesperian ridged
plains material in eastern and southwestern Elysium Planitia and polygonally
grooved material west of the Elysium rise [4]. The extent of the flows is
more clearly defined on Viking than on Mariner 9 images [5], but in some
areas, such as northeast of Hecates Tholus, the flows and smooth plains
material appear to intergrade. We divided the area of the unit into 27
5° X 5° sections each covering 65,000 to 75,000 km'^. The average count
(Table 1) indicates that the flows are lowermost Amazonian, and variance in
the counts allows placement of many of the flows in the Upper Hesperian,
Individual counts range from N(2)=278±61 to 508±84 (craters >2 km diameter per
10° km^), and only three areas have crater densities whose standard-deviation
limits do not overlap with the standard-deviation limits of the average
count, Elysium Mons itself has a crater density of about N(2)=350, which
corresponds to a slightly younger stratigraphic position than indicated by
previous N(l) counts [3,6], Because of the uniformity in crater densities and
the fair to poor quality of images in some areas, we could not distinguish
flow sequences of different periods in most cases. One exception is south of
Eddie crater, where its ejecta have prevented the burial of precrater flows by
younger flows.
The morphology of the flows of this Elysium sequence is about the same as
those of the widespread flow units (members 1 to 5) of the Tharsis Montes
Formation [7], which are Late Hesperian to Middle Amazonian in age [1], The
generally large volumes, lengths, and thicknesses of the flows indicate that
they result from high rates of eruption of low- to moderate-viscosity lava.
Thickness of the flow sequence is a few hundred meters in distal areas, where
they partly bury impact craters several to tens of kilometers in diameter that
were formed on older surfaces [8]; near Elysium Mons, where such impact
craters are apparently completely buried, flows are perhaps more than a
kilometer thick.
The flows mostly originated from the northwest-trending fractures of
Elysium Fossae and fractures circumferential to Elysium Mons, Some fractures,
perhaps where eruption rates were particularly high [9], developed into
sinuous rilles or depressions. These features are mostly found west of
333
Elysium MonSj along with small domes s eroded flows » and crenulated ridges.
The ridges and some of the domes may be eruptive features of high -viscosity
lava. Other domes have summit craters and resemble cinder cones « These
domes, as well as the eroded flows ^ may be composed of pyroclastic material
resulting from the interaction of ground ice with erupting magma.
High up on Elysium Mons, lava flows are shorter^ narrower^ and less
common than those in lower areas. Their smaller size indicates that they were
erupted at lower extrusion rates. They resemble flows that are common high on
Olympus Mons and Tharsis Montes, Also found on Elysium Mons are sinuous
ridges that generally trend downslope; they are prominent on the north flank
of the shield. Other sinuous ridges interfinger with flows south of Elysium
Mons and west of Albor Tholus. Their origin is uncertain, for they resemble
neither wrinkle ridges (of structural origin), nor ridges (of lava-flow or
debris-flow origin) having collapse pits and channels on their crests.
Perhaps the sinous ridges formed by eruption, flow, and erosion of pyroclastic
material.
Also apparently associated with this stage of volcanism are vast eroded
and channelized flows in Utopia Planitia [10]. Their origins have been
ascribed to debris flow [11] and pyroclastic flow [10] caused by volcano-ice
interactions.
Stage 3: Rille volcanism at Elysium Fossae and Utopia Planitia
After an apparent hiatus in volcanic activity, several large rilles of
northwesternmost Elysium Fossae extruded lavas that flowed northwestward into
Utopia Planitia, Morphologies of these flows range from voluminous sheet
flows to thin, narrow flood lavas. Crater counts (Table 1) are approximate
because the flows are not clearly distinguishable from older flows, and areas
proximal to the rilles have fewer craters. Also, age relations with Utopia
flows are unclear; some of the thicker Utopia flows originating from Elysium
Fossae appear coeval with the lava flows. These Utopia flows themselves may
have a volcanic origin related to this period of activity. Given these
relations and the crater counts, we tentatively place these lavas and Utopia
flows in the Lower and Middle Amazonian Series,
Stage 4: Flood-lava and pyroclastic eruptions at Hecates Tholus and Elysium
Mons
A virtually uncratered area west of the summit of Hecates Tholus was
postulated to be mantled by a pyroclastic air-fall deposit [12], The lack of
craters suggests an Upper Amazonian position. Similarly, on Elysium Mons, an
oblong area 60 x 90 km surrounding the summit caldera and extending to the
north and west appears to be mantled by a thin, low-albedo deposit (on images
taken at low sun-incidence angles). The edge of the deposit is composed of
fingerlike extensions radiating downslope. We speculate that this deposit is
composed either of thin pyroclastic flows or air-fall deposits that have
subsequently moved downslope along talus chutes.
On the west flank of Elysium Mons are several patches of sparsely
cratered, ^^ery young flood lavas. Although they resemble the flows described
by [9], we distinguish three separate areas, two of which occur outside the
area mapped by [9], Similarly, most of the youngest flows of Olympus Mons and
Tharsis Montes appear to be thin flood lavas. Very young lavas also have been
postulated to make up smooth plains in the southern part of the Elysium region
[13]; however, we believe that this area is mostly covered by channel material
[10,14].
334
Discussion
Tectonic and channeling activity in the Elysium region is intimately
associated with volcanism [9jl0sl4], Recent work [8jl5] indicates that
isostatic uplift of TharsiSs loading by Elysium Mons, and flexural uplift of
the Elysium rise produced the stresses responsible for the fracturing and
wrinkle-ridge formation in the region. Coeval faulting and channel formation
almost certainly occurred in the pertinent areas in Stages 2 to 4» Older
faults east of the lava flows [8] and channels on Hecates Tholus may be coeval
with Stage 1. Also^ these stages show that the viscosity of erupted magma has
decreased over the eruptive history of Elysium, as it has at Tharsis, perhaps
indicating that the magma originated from progressively deeper sources.
REFERENCES
[1] Tanakas
Mai in,
Plescia
Downs,
Scott,
Neukum,
Scott,
Plescia
Mougi ni
Tanaka
Christ
Mougi n
Plesci
Tanaka
Hall
[2]
[3]
[4]
[5]
[6]
C7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
CITED
K,L, (in press) PLPSC 17.
M.C, (1977) GSA Bull,, v. 88, p, 908.
J.B. and Saunders, R.S, (1979) PLPSC 10, p, 2841.
G.S, and others (1982) J6R, v, 87, p. 9747.
D.H. and Allingham, J.W. (1976) USGS Misc. Inv. Ser. Map 1-935.
G, and Hiller, K. (1981) JGR, v, 86, p. 3097.
D.H, and Tanaka, K.L. (in press) USGS Misc, Inv, Ser, Map I-1802A.
J.B, (1986) LPSC 17 (abstracts), p, 672.
s-Mark, P.J. and others (1984) Earth Moon Planets, v. 30, p. 149.
K.L, and Scott, D.H. (1986) NASA TM 88383, p. 403.
lansen, E.H, and Greeley, R. (1981) LPSC 11 (abstracts), p. 138.
is-Mark, P.J. and others (1982) JGR, v. 87, p. 9890.
a, J.B. (this volume),
K.L. and Scott, D.H. (1986) LPSC 17 (abstracts), p. 865.
J.L, and others (1986) JGR, v. 91, p. 11377-11392,
Table 1.
Crater Densities and Stratlgraphic Positions of Elysium Volcanic Units
Unit
Hecates Tholus mantle
Elysium Mons mantle
Elysium flood lavas
Elysium Fossae-
-Utopia Planitia flows
Elysium Mons flows
Nixl
Crater density
= no. > X km d1am./10° km*^
no fresh craters observed [12]
few kilometer-size craters
few kilometer-size craters
N(2) = about 100-200
Stratlgraphic series
Upper Amazonian
Upper Amazonian
Upper Amazonian
Lower Amazonian-
Middle Amazonian
all
N(2) = 378±15
Upper Hesperian-
shield
N(l) = 2350+153 [3]
N(l) = 1800-4800 [6]
N(2) = about 350
Lower Amazonian
Albor Tholus
N(l) = 1500±263 [3]
Upper Hesperian
Hecates Tholus
N(l) = 1800+351 [3]
N(l) = 1300-15,000 [6]
Upper Hesperian
Note: Stratlgraphic series determined by crater counts and stratlgraphic
relations (see text).
335
LATE-STAGE FLOOD LA¥AS IM THE ELYSIUM SEGIOM, MMS J,B. Plescia, U.S.
Geological Survey, 2255 N. Gemini Drive, Flagstaff, AZ 86001
In the southeastern part of the Elysium regionj centered near lat 5° N.,
between long 180° and 210°, and adjacent to the Cerberus Rupes fractures, is a
unit that exhibits little texture and a generally low albedo and that has a
very low crater frequency (Figure 1). This unit has been mapped as "smooth
plains material" and interpreted as an eolian deposit on the basis of Mariner
9 images (Scott and Allingham, 1976). More recently, Scott and Tanaka (1986)
mapped the unit as material deposited during a channeling episode. Here,
however, I interpret the smooth plains unit as being a volcanic deposit
composed of low-viscosity lava flows; both flood lavas and individual flows.
In images that have resolutions of 200-300 m/pixel, the unit is
characterized by a smooth texture and albedo patterns suggestive of flowage in
an easterly to northeasterly direction. In higher resolution images (30-40
m/pixel) , the large-scale albedo patterns in some locations are recognized to
be lava flows having well-defined lobate margins. This surface represents the
youngest widespread unit in the Elysium region and one of the youngest
volcanic units on the planet (Table 1).
The plains formed by these young volcanics are locally interrupted by
inselbergs of knobby material and by kapukas of a brighter material. The
brighter material typically lies adjacent to dark, narrow channels or forms
teardrop-shaped islands within the channels. In the high-resolution images,
the brighter material is seen to be an older, more cratered unit that has been
channelized, the channels being filled with a dark, smooth material — the
volcanic flows.
The volcanics flowed through preexisting channels eastward and
northeastward into and across the knobby terrain east and west of Orcus Patera
and continued to the north into western Amazonis Planitla (Figure 1). East of
Orcus Patera, the volcanics flowed through well-defined channels whereas west
of Orcus Patera, it moved through a myriad of narrow channels. This region is
mapped in Figure 1 as Volcanic plains, mixed (vpm) because individual flows
cannot be differentiated from older units at this scale.
The volcanics apparently have a source around Cerberus Rupes as indicated
by the flow pattern, and several vents have been recognized in the
southwestern part of the region (Figure 1). The vents are poorly resolved but
appear to be low shields having a central vent surrounded by radial flows.
All of the vents occur in areas where the lavas are thin, as indicated by the
degree of exposure of preplains craters and the frequency of inselbergs.
These vents may reflect either a late-stage eruptive style or one specific to
this area. Elsewhere, where such vents have not been observed, the lavas may
have erupted through fissures ultimately buried by the eruptions. Although
centrally located, Cerberus Rupes can not be confidently considered a vent
because flows have not been observed to originate along it. In addition.
Earth-based radar-topography profiles across Cerberus Rupes suggest that it is
a series of normal faults having several hundred meters of displacement; hence
it is probably a late-stage tectonic feature.
Tanaka and Scott (1986) studied this same region and came to a different
conclusion regarding the nature of the smooth, young material. They
recognized the channeling aspects of the region, but attributed the origin of
336
the unit directly to the channeling episode; hence they interpreted the young
age of the unit as indicative of a major late-stage fluvial episode. In
contrasts I interpret these plains to be volcanic flows which locally fill
preexisting channels. ThuSj the channels represent an older, probably
unrelated event. The source region for the channeling fluid is unknown;
perhaps it is now buried by the volcanic plains, or perhaps it lies to the
south beneath material interpreted to be ash flow and ignimbrite material
(Malin, 1979; Scott and Tanaka, 1982),
This volcanic unit is important for several reasons » Its presence
indicates that in Elysium, unlike in Tharsis, plains volcanism continued or
resumed after shield building ended. Additionally, this late-stage volcanism
was of sufficent volume and low viscosity that it covered a wide area. The
lack of topography, even in high resolution, and the distance the flows
traveled suggest the material had a very low viscosity. Only at its distal
end, in Amazonis, are flow features and lobate fronts well developed. This
material further indicates that there was sufficent heat in the upper mantle
to generate considerable quantities of "basaltic" lava at a high eruption rate
late in Martian history.
REFERENCES: Malin, M.C, 1979, Mars: Evidence of indurated deposits of fine
materials: NASA Conf, Pub. 2072, p. 54; Scott, D.H., and Tanaka, K.L., 1982,
Ignimbrites of Amazonis Planitia region of Mars: J. Geophys. Res., v. 87, p.
1179-1190; Scott, D.H., and Allingham, J.W., 1976, Geologic map of the
Elysium quadrangle of Mars: U.S.G.S, Misc. Inv, Series Map 1-935; Tanaka,
K.L., and Scott, D.H., 1986, The youngest channel system on Mars: Abstracts
Lunar and Planet. Sci, Conf. 17th, p. 865-866.
TABLE 1
CRATER FREQUENCIES
FREQITENCY OF CRATERS 2. D / 10^ km^
DIAMETER
TERRAIN I iM 2k!a Sim 10 km
Knobby terrain (kt)
Ridged plains (rp)
Elyisum plains (Epu)
Volcanic plains (vp)
2780+275
1580+205
790+145
390+100
2730+190
780+100
252+ 60
80+ 30
1420+ 74
440+ 40
90+ 20
35+ 10
90+ 15
23+ 8
5+ 4
3+ 3
337
CO
CO
00
•'/ _ . . _
///////////////////^^
''/////////^///y///////
^/if//yy //////// //////
<'///////////////////////
////////////////////////
////////////////////////.
////////////////////////.
///////////////y////////
B cJ^// /y /////// y //////// /
' 777/ /y/ //////////////// /
////////////////////////
////////////////////////
////////////////////////
////////////////////•////
////////////////////////.
////////////////////////
////////////////////////
r///////////////////////
////////////////////////
////////////////////////
y///////////////////////
////////////////////////
////////////////////////
/////////////////////yy/
/////'/////////////////yy
///////////////yy//////y
/////////////////y/A////
/////////////////// * ■ ■ ■ ■
EXPLANATION
^H Volcsnic plstins
//////
//////////
//////////
//////////////
/////////////////
///////////////////
///////////////////
////////////////////////
_. ///////////////////////////
^J^// ///////////////////////// // "
J^ /////////////////////////// ///
f///////////i'////////////////////
^v ////////////////////////////////
3/////////////////////////////////
f/ ///////////////////////////////// ,
_Y/ /////////////////////////////////
f? ///////////////////////////////////
V ////////////////////////////////////
/////////////////////////////////////
f////j'/////y//////////y////////////////
y///j'////////////////y/////////////////
,. - . ////// ////yyyy///y /yyyy ///'////'/////////
y//////////////////////////////////////////
Flow Indicator
Wrtnkitt r\tie9
Fauti
W Elysium volcsnic
^^ units undlffersntlstttd .
Rldgod plains
and corr«tetiv» units
Knobby t^rrafn
Undlff«rentlffit9d terrain
206* W
180* W
FIGURE 1.
Geologic map of southeastern Elysitim Region, Mars,
ielativ® ^i®s Of Lwa flms M MM Patera, Mars
by
Dale M. Schnaebergar and David C. Piari
Jet Propulsion Laboratory, California Institute Of Technology, Pasadena. CA
Many Wgs lava flows on the flanks of Alba Patera are astonishing in their volume and length
(11). They are enormous by terrestrial stancferds ranging from 60 kilometers to over 500 kilometers
m length. As a suite, these flows sug^t tremendously voluminous and sustained eruptions, and provicte
dimensional boundary conditions typically a factor of 1 00 lar^r than terrestrial flows ( 2). One of the
most striking features asssKiated with Alba Patera is the lar^, ralially oriented lava flows that exhibit
a variety of flow morphologies. The^ include sh^t flows, tube- fed and tube-channel flows, and
undifferentiated flows ( 3 ,4 ,5 ,7).
Thrffi groups of flows were studied; flows on the norhtwest flank, southeast flank, and the
intracaldera r^ion. Flow F4 is part of a system of complex flows located on the northwest flank of Alba
Patera. The two sheet flows (F5 and F6) on the southeast flank appear to be less complex as compared to
tho^ on the northwest flank but re^mbles the lobate flow morpholi^ of flow ?A. Within the confines
of the concentric fra:ture/graben system lies the central caldera complex with its intracaldera flows.
Six flows (F7 to F12) were ictentified here and, although their dimensions are consicterably smaller,
they continue to exhibit lobate morphology.
Two general models have b^n proposed (10,14) relating observed crater densities to absolute &^
using crater diameters of either > 1 Km or >4 Km. A review of the literature uncovers that a number of
crater ctenslties and absolute ages have been reported for Alba Patera (9,12). Within the framework of
these models their data suggests a maximum absolute age for the Alba shield of 3.8-3.5 bybp ( 1 ,9 , 1 0)
down to a minimum age of 2.6- 1 .0 bybp (1 ,9, 1 4). Plescia and Saunders ( 1 979) ( 1 2) have su^staJ
that a transition in the types and loci of cone building volcanism occurred from about 2.0 to 1 .5 bybp
beginning at the time of major activity at Alba Patera (12). In contrast, Neukum and Wise
( 1 976 )( 1 0) believe that by 3.0 bybp the major tectonic/volcanic disturbances were winding down and
that by 2.5 bybp the activity was essentially over. These differences remain unresolved.
The above cited ages are for r^ional studies of the entire Alba shield whereas our work concernaj
itself with attempting to deduce the ages of the individual flows. Plotting of cumulative size-frequency
distribution curves of the flows shows a general clustering of the data, except for flows F6 and F7 which
are noticeably shifted vertically from the rest (Figure 1 ). The apparent clustering of most of the flows
is an artifa:t of the overlapping of confictence Intervals as determined from an assumed Poisson
distribution ( 8, 1 3, 1 5) of the data. Crater (tensity data was also plotted using crater diameters binnaj
in /2 Intervals ( 15). These data show a clustering of the flows similar to the cumulative curves with
flows F6 and F7 being somewhat distinct from the rest (Figure 2).
Crater diameters measurKl for the flows on the flanks and intracalctera region ran^ from about
100 meters to 2.62 kilometers. However, the majority measured (387 out of 41 1 ) are less than 1
kilometer In diameter. The u^ of a 1 kilometer datum for the determination of both relative and
absolute ages has been widely u^ In the literature (3,6,10,13), Using this datum, two widely
^paratKl e^es were determined frwn the cumulative size-frequency plots (Fig 1 ). Plotting this date
against the Neukum and Wise ( 1 975) ( 1 0) curve yielcfed an absolute age range of 3.9-3.3 bybp while
plotting the same data a^inst the Soderblom et ai. ( 1 974) ( 1 3) curve yielded a range of 0.8-0. 1 bybp
(Fig 3). Craters numbers, independent of size-frequency plots, suggest an absolute sgs ran^ betwe^i
1 .3 and 0.4 bybp after calibrating the data to fit the derived martian ab^lute age curve (for craters >4
Km) by S«ferblom et al. ( 1 974) ( 1 7).
A relative sgs sK^uencs can be deducKl from size-frequency distribution plots. Cumulative
siffi-frKiuency distribution curves suggest that flows F6 and F7 are distinct from the apparent
clustering of the other flows. The remaining flows (F4 , F5 , F8 , F9 , F 1 , F 1 1 , and F 1 2) have plots that
are very close with consitferable overlap of their confictence intervals. It ^ms unlikely that any
confident separation of relative ages for these flows is possible.
Like the cumulative curves above, the data arrangaj in •/2 diameter bins to produce
size-frequency distribution curves shows that flows F6 and F7 appear distinct from the others with flow
F7 being the olifest. Again, a similar clustering of flows into an intermajlate group is su^stej. Flows
F8 , F 1 1 , and F 1 2 appear to be lost in the noi^ typical for the smaller crater diameters less than 1 Km ,
339
althcm^ aimewhat better ^parated than in the cumulative curves, if the vertical displacements of
curves are real, the su^iestKl relative ages of the flows is F7 (oldest), followKl by flows F 10, F9, F5,
F4 as intermKJiate, and flow F6 as the youngest. The relative a^ positions of flows F8, F 1 1 , and F 1 2
remain , for the most part , uncteterm insd.
Relative a^ based on crater numbers, for craters i 1 km intercept, suggest a ^uence, olctest to
youngest, of F7,F4,F10,F9, F5, andF6. Flows F8, F9, Fll , and F 12 are undetermined because they
lack craters i 1 km diameter. The smaller diameter crater numbers, of probable secondary origin,
show a somewhat different sequence than above. Both cfeta sets agree that flow F7 is the oldest and F6 the
youngest, in contrast, however, the latter suggests that flow F4 is younger than both flows F 1 and F9.
Additionally, it allows for the placement of two of the three previously undetermined flows ( F8, F 11 ) on
either side of flow F4, older and younger, respKtively. The relative age of flow FI2 still remains
undetermined. Since two possible age sequences are suggested, depending on which data set is accepted as
representative, a relative age ^uence based on both data sets can be made so long as the larger diameter
( i I Km) data is given more weight. Thus a possible sequence, from oldest to youngest, is F7, F4, F 1 0,
F9, F8, F 1 1 , F5 and F6. Perhaps flow F 1 2 is the youngest of all as it has the least number of craters.
However, since there are no craters in either size bin, judgement as to its age will not be attempted.
Source vents, as a rule, are not clearly defined on the flanks of Alba Patera. However, some
estimation of vent sequence can be macfe. The data suggests that the earliest vent activity was from the
northeast part of the caldera complex producing flow F7. This area appears to have had recurring
activity at later times as well to produce flows F 1 and F9. The southeastern part of the caldra complex
also became active producing lavas that fed flow F5. Perhaps simultaneous with this activity (within the
scope of crater ages), lavas erupted from vents or fissures of unknown location on the northwest flank
to produce flow F4. It is possible that the northwest flank flows may have erupted from a lateral vent on
the patera's flank at or near the present ring fracture zone. Finally, the southeast part of the caldera
became active once again to produce flow F6 (or possibly F5). Flows F8, Fl 1 and F12 were erupted
from the northern half of the caldera most likely sometime after the effusion of flow F7 but before or
simultaneously with flows F 1 , F9 and F5. This suggests that the vents and/or fissures associated with
the effusion of lavas from northern half and southeast portions of the caldera complex were perhaps
operating within a relatively narrow geologic time frame. Unfortunately, high resolution delineation of
vent activity beyond a rudimentary relative sequence is not possible with current crater density
analysis techniqies.
In summary, the lava flows discussed above probably were erupted as group during the same
major volcanic episode as suggested by the clo^ grouping of the data. Absolute dgss are poorly
constrainsj for both the individual flows and shield, due in part to disagreement as to which absolute age
curve is representative for Mars. A relative age sojuence is implied but lacks percislon due to the
closeness of the size-fraiuency curves. R^rdless, it appears that the final stages in Alba's volcanic
history were anything but quiet.
REFERENCES
( 1 ) Basalttc Volcanism Study Project ( 1 98 1 ) New York. Perqammon Press. 1 286p.
(2) Bologa, S. M., and Pieri, D. C. (1985) NASA TM-87563. p. 245-247
(3) Cm-r. M. H. (198 1 ) The Surface Of Mars: Hew Haven. Yale University Press. 232 p.
(4) Cwr. M. H.. el al. (1977) Jour. Geophvs.Res.. V.82. p. 3985-4015
(5) Carr. M. H., and Greeley. Ronald (J980) NASA SP-403
(6) Dial, A. L. (1978 ) NASA Th 79729. p. 179-181
(7) Greeley. Ronald, and Spudis. P. D. (1981) Rev. Qeophvs. Space Phvs. 19. p. 13-41
(8) Hartmann. W. K. (1973) Jour. Geophvs. Res.. V. 78. p. 4096-41 16
(9) Hitler. Konrad. and Neukum, Gerhard (1980) NASA Tech. Memo. 81776. p. 1 19-121
(10) Neukum. 6.. and Wise. D. U. (1976) Science. V. 194. No. 4272. p. 1381-1387
(1 1) Pieri. D. C. et al. (1986) NASA TM 88583. p. 318-320
(12) Plescia, J. B.. and Saunders. R. S. (1979) Proceedings. LPSC-X. p. 2841-2859
(13) Soderblom, L. A. et al. (1974) Icarus 22. p. 239-263
(14) Soderblom. L. A. (1977) m Roddy. D. J., Pepin. R. 0., and Merrill, R. B.. eds., Impact and
Exploration Cratering: Perqamon Press. New York, p. 629-633
(15) Woronow, A., et al. ( 1978 ) NASA TM 79730. 20 p.
340
LogCN) 4
Craters
2 ■
45,000
•«0,000
J5,000
30.000
CumuUtiv«
Cral*r Fr»qu«nc<j
NClO-^Km-Z) ^'-"^
For Craters ilKm
20,000
10,000
5,000
N»uHim8.ViseC1976)
SoOsrtilimi ♦til (1974)
Log Diameter (Km)
4 3 2 1
AGE(IO*Yfjrs)
Figure 1
Figure 3.
5 T
4
U>B«of
|MrlO-6
Kmn 3
2
-1.0
-05
0.0
Lsgcr^ardiiiMUr.
^2 IncrwMntbfn
«-
H
0-
F5
e-
F6
o
F7
^
FB
A-
F9
X-
FIO
ac
Fll
-~
FI2
03
Figure 2.
341
MARTIAN VOLCANISMj FESTOON-LIKE RIDGES ON TERRESTRIAL
BASALT FLOWS AND IMPLICATIONS FOR MARS
E. Theilig and R. Greeley, Dept. of Geology, Arizona State University, Tempe,
Arizona, 85287
Determination of lava flow compositions on Mars is a significant problem with major
implications for the thermal history and differentiation of the planet. SiUca content within a
magma is one of the factors controlling rheological properties of a flow. Thus, one of the
major focuses within martian volcanology has been to estimate the composition of a lava
flow based on rheological properties determined from flow morphology. One method,
derived by Fink and Fletcher [1] and Fink [2], which has been used is the analysis of
regularly spaced, arcuate, festoon-like ridges oriented perpendicular to flow direction.
Their model relates ridge height and spacing to lava rheology, thickness of the flow's
thermal boundary, and applied stresses and allows the viscosity of the interior of the flow
at the time of ridge formation to be estimated. Festoon ridges on martian lava flows are
similar in size to those on terrestrial silicic flows and previously have been compared to
rhyolitic, dacitic [2-4], and trachytic [5] flows. In this study we use the Fink and Fletcher
[1] and Fink [2] model to assess and compare flow rheology for two terrestrial basalt flows
and one martian flow with previous studies.
Lava flows selected for this study include the Lakagigar and Barthardalshraun flows,
Iceland and a flow west of Arsia Mons on Mars, located at 3°S, 138.2°W. The Lakagigar
flow [6], more commonly referred to as Laki, contains festoon ridges which occur locally
but which are dominant where the flow spreads out on a coastal plain. Festoon ridges on
the Barthardalshraun lava flow, Iceland are located in an area of the flow which ponded in a
valley ~ 80 km from the vent [7]. On the martian flow west of Arsia Mons, festoon ridges
occur across large flow lobes [4]. Values of ridge spacing and height used for this study
are shown in Table 1. Statistical analyses of the data indicate that a dominant spacing exists
for the flows included in the study [8]. This suggests that a strong folding instability
existed, thus the model of ridge growth should be applicable to these flows.
Fink and Fletcher [1] and Fink [2] considered festoon-like ridges to be folds resulting
from compression of a fluid in which viscosity decreases with depth. Regular ridge
spacing indicates a folding instability which places constraints on dimensionless parameters
expressing the ratio of surface to interior viscosity, the ratio of gravitational stress to
compressive stress, and ridge spacing [2]. Based on these constraints, minimum interior
viscosity at the time of ridge formation can be estimated from ridge height and spacing,
determined either in the field or from image data [2,3,5]. The dimensionless groups used
for this analysis are:
R = rio/Tli (1)
InR > 30 h/d (2)
erii>pgh/(0.08RlnR) (3)
where R is the ratio between the exterior viscosity (Tjo) and interior viscosity (j\{), h is the
thickness of the thermal boundary layer approximated by ridge height, d is the average
ridge spacing, € is the strain rate, p is lava density, and g is the gravitational acceleration.
To estimate the minimum interior viscosity, strain rate can be approximated from finite
strain and an assumption of ridge growth time.
342
Results are shown in Table 2 and compared with previous studies. Estimated
viscosities for the Icelandic flows are high (5x10^ - SxlO^l Pa s) for terrestrial basalts
which typically range from 10^^ - 10* Pa s [9]. Higher than normal viscosity can be
obtained, however, by decreasing temperature, increasing solid content in the magma, or
decreasing gas content, all of which are related. In the cooled, terminal areas of basaltic
flows on Mount Etna, viscosities reached 10^ - IQlO Pa s as the flow halted [10]; and for a
basalt flow on Mauna Loa, viscosity was estimated at 10''' Pa s at the toe [11]. Cigolini et
al. [12] determined viscosities of ~ 10^ Pa s from both field measurements and
experimental results for basaltic andesite flows on Arenal Volcano, Costa Rica. They
attributed the high viscosity values to a high crystal content in the magma and low effusion
temperatures. Thus the viscosities for the Icelandic flows indicated by the formation of
festoon ridges are not unreasonable for mafic magmas. The dominant location of these
ridges on the terminal lobes of the Laki flow and in the ponded section of the
Barthardalshraun flow suggest that cooling was a significant factor in increasing the
viscosity in these flows.
The minimum interior viscosity values are comparable to those for a trachyte flow on
Hualalai, Hawaii (7.6x10^ - 7.6x101'^ Pa s), a flow on Ascraeus Mons (1.1x10^ -
l.lxlOlO Pa s) [5], a dacite flow in Chile (> 4.6x10^ Pa s), and some ridges in Arcadia
Planitia (10^ Pa s) [2]. The similarity in results (Table 2) probably reflects a requirement
for lavas to have high viscosities before this size ridge will form. Because basalt may have
a high viscosity under specific conditions ridge height and spacing may not represent
compositional variations. Thus, caution should be used in applying this model to obtain
rough estimates of composition.
Based on the morphologic similarities between the martian flows and the Icelandic
flows and knowledge of the emplacement of the terrestrial flows, the flows west of Arsia
Mons are considered to have been emplaced as large sheet flows from basaltic flood-style
eruptions. Festoon ridges represent folding of the surface crust in the last stages of
emplacement when viscosities would be high due to cooling. Alternatively, the lava may
have had a high crystallinity or was erupted at low temperatures. In addition, increased
compressive stress behind halted flow fronts or in ponded areas may have contributed to
ridge formation.
REFERENCES
[1] Fink, J. H., and R. C. Fletcher, Ropy pahoehoe: Surface folding of a viscous fluid, /.
Volcanol. Geotherm. Res., 4, 151-170, 1978.
[2] Fink, J., Surface folding and viscosity of rhyolite flows. Geology, 8, 250-254, 1980.
[3] Fink J. H., Possible rhyolite flows in the Arcadia Planitia region of Mars: Evidence
from surface ridge geometry (abstract), in Lunar and Planet. Sci. XI, 285-287, Lunar
and Planetary Institute, Houston, 1980.
[4] Schaber, G. G., Radar, visual and thermal characteristics of Mars: Rough planar
surfaces, /carMs, 42, 159-184, 1980.
[5] Zimbelman, J. R., Estimates of rheologic properties for flows on the Martian volcano
Ascraeus Mons, /. Geophys. Res., 90, D157-D162, 1985.
[6] Thorarinsson, S., The Lakagigar eruption of 1783, Bull. Volcanol., Ser. 2, 33, 910-
927, 1970.
[7] Greeley, R., and H. Sigurdsson, Pristine morphology of a quasi-flood basalt flow:
The Bardardalshraun of Trolladyngja, Iceland (abstract), in Reports Planetary
Geology Program 1980, NASA TM 82385, 245-246, 1980.
343
[8] Theilig, E., and R. Greeley, Lava flows on Mars: Analysis of small surface features
and comparisons with terrestrial analogs, /. Geophys. Res., 1986, in press.
[9] Basaltic Volcanism Study Project, Basaltic volcanism on the terrestrial planets, 1286
pp. Pergamon Press, Inc, New York, 1981.
[10] Walker, G. P. L., Thickness and viscosity of Etnean lavas. Nature, 213, 484-485,
1967.
[11] Moore, H. J., Preliminary estimates of the rheological properties of the 1984 Mauna
Loa lavas, U.S. Geol. Surv. Prof. Paper 1350, 1986, in press.
[12] Cigolini, C, A. Borgia, and L. Casertano, Inter-crater activity, aa-block lava,
viscosity and flow dynamics; Arenal Volcano, Costa Rica, /. Volcanol. Geotherm.
Res., 20, 155-176, 1984.
Table 1. Ridge Geometry Data Used in Calculations.
Study Area
Values Used in
Calculations
Ridge Spacing
Height
(ml
Spacing
(m)
Range
(m)
Number of
Samples
Median
(m)
Mean
(m)
Standard Dispersion
Deviation
West Arsia Mens
12
90
30-194
120
89.0
91.7
29.7 0.3
Laki Flow
Interior
1
2
2
10
10
15
7-23
56
13.1
12.9
4.3 0.3
Margin
2
16.5
9-37
65
16.4
16.8
5.3 0.3
Barthardalshraun
Flow
5
100
—
~~
^
—
"— —
Table 2. Estimates of Interior Viscosity for Martian and Terrestrial Lava Flows
Study Area
Height Spacing
en
Interior Viscosity (Pa s)
(m)
(m)
IHour
IWeek
West Arsia Mons
12
90
6.5 E3
1E8
2E10
Laki Flow
Interior
1
10
5.1 E3
1E8
2E10
2
10
2.5 E2
5E6
9E8
Margin
2
16.5
4.4 E3
1E8
2E10
Barthardalshraun Flow
5
100
2.3 E5
5E9
8E11
Chao Flow (Dacite)*
30
100
5E8
Arcadia Planitia*
100
750
1E8
1E9
Hualalai (Trachyte)t
15
150
8E8
8E10
Ascraeus Monst
15
120
1E8
lElO
* [2] T[5]
"E" Indicates the expon(
;nt in pow
ers of 10.
344
STUDIES OF FLUID INSTABILITIES IN FLOWS OF LAVA AND DEBRIS
Jonathan H. Fink, Geology Department, Arizona State University, Tempe, AZ 85287
Introduction
Estimating the rheology of lava flows is an essential tool for the remote
determination of flow compositions on other planets and is also a component in the
evaluation of many volcanic hazards. For flows whose emplacement is not observed,
indirect methods must be used to assess such physical parameters as viscosity, yield
strength, density, and volatile content. Most studies of this type have assumed that the
geometry of large scale morphological features on the flow surface reflects the bulk
rheology of the active lava. One class of lava surface structures that is particularly
well-suited to this sort of interpretation includes those periodic features that result
from fluid instabilities.
At least two instabilities have been identified and utilized in lava flow studies:
surface folding and gravity instability. Both lead to the development of regularly spaced
structures on the surfaces of lava flows. The geometry of surface folds, first analyzed
by Fink and Fletcher (1978), has been used by Fink (1980a), Fink et al. (1983), Zimbelman
(1985), Head and Wilson (1986), and others to estimate the rheology of lava flows on other
planets. Fink and Fletcher's (1978) analysis assumed that lava flows have a temperature-
dependent newtonian rheology, and that the lava's viscosity decreased exponentially inward
from the upper flow surface.
The presence of a gravity or Taylor instability was proposed by Fink (1980b) to
explain certain regularly spaced domal outcrops of low density pumice on the surfaces of
silicic lava flows. Subsequent investigations (Fink, 1983; Fink and Manley, 1986; Manley
and Fink, submitted) have attempted to relate the density inversion to the distribution
and migration of volatiles within actively advancing flows. Such studies provide possible
criteria for the identification of silicic lava flows on high resolution images of other
planets, since diapirism results in flow surfaces with very high albedo contrasts, and the
associated concentrations of volatiles may lead to formation of large craters and other
explosive features on the surface of a silicic flow. These models are also important for
hazards studies since the presence of volatiles in silicic magmas is considered a major
factor in the inception of explosive volcanism.
Analyses of surface folding
Surface folding (Biot, 1960; Fink and Fletcher, 1978) produces regularly spaced ridges
that range from cm-scale ropes on pahoehoe basalt flows to 100-m-scale lava ogives seen on
rhyolite flows. Analysis shows that ridge spacings depend on the thickness of the cooling
crust, magnitude of the compressive stress, and weight of the lava. Ridges can only form
when the ratio between surface and interior lava viscosities lies in a restricted range:
if the crust is too stiff (cool), the surface deforms by fracture rather than flow; if too
soft (hot), ridges form but are not retained.
345
The application of this model to extra-terrestrial lava flows has been attempted by
numerous investigators. One of the main drawbacks of the model has been that it requires
estimating the thickness of a flow's thermal boundary layer (crust). Previous planetary
applications have required assumptions about the strain rates and times of formation
associated with large ridges (formation times have been assumed to be between 1 day and
one week), and about how the crustal thickness relates to the amplitude of the ridges
(crust thickness was generally assumed equal to ridge amplitude).
A lower bound on strain rates was obtained from observations of the active 1984 Mauna
Loa basalt flow. During this eruption, surface ridges with wavelengths ranging from 10-30
m and amplitudes of 1-2 meters were seen to form relatively abruptly at positions that
migrated upstream as part of an overall change in aa surface characteristics described by
Lipman, Banks, and Rhodes (1986). Ridges were seen to form behind flow constrictions over
a period of less than a day. The observed wavelengths and amplitudes correspond to
strains of between about 0.003 and 0.080. If these occurred over periods of less than a
day, the minimum strain rates would be between 10"^ and 10"^ sec"\ which are considerably
less than those used in earlier calculations. Lower strain rate estimates in turn lead to
higher calculated viscosity values.
Analyses of Taylor instability in lava flows
Recent drilling investigations in rhyolite flows (Eichelberger et al., 1985; Goff et
al., 1986) have provided new information about their volatile distributions, and now allow
more accurate specification of boundary conditions for the gravity instability model.
Fink's original linear analysis (1980b) assumed newtonian rheology, an absence of
slope-induced shear stresses, a rigid upper flow surface, and a two-dimensional
perturbation of the unstable interface. The new drill data have motivated a more rigorous
and generalized analysis of the Taylor instability in lava flows.
The new formulation allows for three dimensional disturbances, the effect of a mean
shear flow on the instability, and a deformable free upper surface. This analysis allows
determination of whether transverse ridge formation is favored over other possible
orientations, assessment of the deformation of the free surface relative to that of the
unstable internal interface, and use of the viscosity ratio to help determine the state of
volatiles within the buoyant layer.
Results of the new analyses (Baum et al., in review) include calculations of preferred
wavelength and growth rates as functions of buoyant and dense layer thicknesses and
viscosities. In general, diapir spacing increases in proportion to the thickness of the
buoyant layer and to the viscosity of the dense layer. The measured thicknesses of buoyant
and dense layers in three different drill cores and in two well-dissected rhyolite flows
in Arizona allow the relative viscosities of the layers to be determined. It was found
that the presence of regularly spaced, dark, pumiceous diapirs on the surface of
terrestrial sihcic flows reflects a local concentration of volatiles of from 3 to 6 times
the average values. Recognition of such diapiric structures in conjunction with unusually
large surface ridges on martian or venusian flows would suggest more evolved lavas than
have previously been identified. Such features are currently being sought in high
resolution Viking images.
346
Surface folding on debris flows
Not all terrestrial flows that preserve regularly spaced ridges are lavas. Debris
flows, landslides, mudflows, and other sorts of mass movements may all develop surface
folds in response to flow parallel compression. As part of a study of the Chaos Jumbles
debris avalanche deposit at Lassen National Park, an analysis of surface folding
instabilities on plastic materials was conducted (Eppler, Fink and Fletcher, 1987). The
main differences between the rheological model used for debris avalanches and for lava
flows were (1) the debris was assumed to have an isothermal, power-law rheology with a
finite yield strength, rather than a temperature-dependent newtonian rheology; and (2) the
debris was assumed to move primarily by slip over a frictionless basal surface rather than
by continuous flowage.
A surprising result of the analysis was that the spacing of ridges was proportional to
the total thickness of the flow, rather than to the crustal thickness alone. The ratio of
ridge spacing to flow thickness ranged from 1.4 to 2.8 for all reasonable rheologic
values. These two endmembers correspond to relatively low and high values of debris
strength. Hence, by measuring the spacing of ridges on a debris flow, it may be possible
to estimate its thickness (and volume) to within a factor of two. For the case of Chaos
Jumbles, volume estimates based on ridge spacing were consistent with volume estimates
based on the size of the collapse scar left behind by the falling debris. Application of
this model to the interpretation of landslide deposits on Mars is currently in progress.
References
Baum, B., Fink, J., Krantz, W., and Dickinson, R., Rayleigh-Taylor instability in rhyolite
flows. JGR (submitted)
Biot,M., 1960, J. Franklin Inst , 270: 190-201.
Eichelberger, J., Lysne, P., Miller, C, Younker, L., 1985, EOS, 66: 186-187.
Eppler, D., Fihk, J. and Fletcher, R., 1987, Rheology and kinematics of the Chaos Jumbles
debris avalanche deposit, Lassen National Park, CA. JGR (in press).
Fink, L and Fletcher, R., 1978, J. Volcanol. Geotherm. Res. , 4: 151-170.
Fink, J., 1980a, Geology, 8: 250-254.
Fink, J., 1980b, Tectonophysics , 66: 147-166.
Fink, J., Park, S., and Greeley, R., 1983, Icarus , 56: 38-50.
Fink, J., 1983, Geol. Soc. America Bull , 94, 362-380.
Fink, J. and Manley, C, 1987, in Fink, J., editor, G.S.A. Special Paper 212, (in press).
Goff, F., Rowley, J., Gardner, J., Hawkins, W., Goff, S., Charles, R., Wachs, D., Maassen,
L., and Heiken, G., 1986, JGR, 91: 1742-1752.
Head, J. and Wilson, L., 1986, JGR, 91: 9407-9446.
Lipman, P., Banks, N., and Rhodes, J., 1986, Nature , B17: 604-607.
Manley, C. and JFink, J., Intemal textures of rhyolite flows as revealed by
research drilling. Geology (submitted).
Zimbehnan, L, 1985, Proc. Lunar Planet. Sci. Conf., 16: D157-D162.
347
LONG, PAIRED A'A/PAHOEHOE FLOWS OF MAUNA LOA: VOLCANOLOGICAL SIGNIFICANCE.
AND INSIGHTS THEY PROVIDE INTO VOLCANO PLUMBING SYSTEMS
Scott K. Rowland &c George P.L Walker, Hawaii Institute of Geophysics,
2525 Correa Rd., Honolulu, Hawaii 96822
Martian lava flows of Olympus Mons and the Tharsis volcanoes are an order of mag-
nitude longer, wider, and more voluminous than any single lava flows on Earth (cf. Carr,
1981). The long lava flows of Mauna Loa, Hawaii have been cited as Earth's closest ana-
logs to the large Martian flows (Carr, 1981). It is therefore important to understand the
flow mechanisms and characteristics of the Mauna Loa flows and to make use of these
in attempts to gain insights into Martian eruptive processes.
A number of Workers (Walker, 1973; Malin, 1980) have made attempts to find regu-
lar relationships between flow characteristics such as flow thickness, length, and area
with physical controls of lava flows such as underl5ring gradient and average eruption
rate. Assumptions regarding lava rheology can then be made if a quantitative m.ethod
of analyzing the found relationships exists.
The careful mapping of selected Mauna Loa flows has revealed the unexpected
relationships that lavas with very different flow mechanisms can have very similar
volumes, areas, and lengths. The division of Hawaiian lavas into pahoehoe and a'a (cf.
Macdonald, 1953) is a fundamental one and is useful for many reasons. The common
down-flow change from pahoehoe to a'a due to degassing, cooling, and shearing (Emer-
son, 1926; Peterson & Tilling, 1980) is well established; what is not so well known is that
flows having similar lengths (up to 47 km) and volumes (up to 200 x 10® m^) can consist
of pahoehoe or a'a in their entirety.
Our mapping has revealed the existence of several paired lavas, each consisting of
an earlier a'a member and a later pahoehoe member erupted in the same eruption, the
two members of a pair having similar lengths and volumes and, so far as we know,
identical initial viscosities and chemical compositions.
The 1859 and 1880-81 lavas of Mauna Loa are particularly striking examples of
paired a'a/pahoehoe flows. Historical accounts allow us to piece together the charac-
teristics of the eruptions, which we can then relate with features of the eruptive pro-
ducts analyzed in the fleld and in air photos. An important relationship which we find
for these two eruptions is the close correlation of erupted lava type with discharge
rate; channel-fed a'a formed at a time of high discharge rate and tube-fed pahoehoe
formed at a time of low discharge rate. Eruptive style is also closely related to
discharge rate; high gas-driven fountaining accompanied the high discharge-rate
activity, and little or no fountaining accompanied the low discharge-rate activity. A
similar relationship was noted on Kilauea during the Mauna Ulu series of eruptions
(Swanson, 1973; Swanson et.al., 1979), and is at present being demonstrated by the
1983-1986 Puu O'o and "C-flssure" (Ulrich et.al.. 1986) activity.
At a time of high discharge, lava is strongly channelized and the high flow velocity
precludes the roofing-over of significant lengths of the main channels. The loss of heat
from open channels is very rapid, and a significant viscosity increase occurs in the
channelized lava. Because of lateral velocity gradients, any portion of chilled lava
crust is subjected to a torque due to traction by the underlying lava, and because of
this torque adjacent portions of skin tend to be torn apart. If the underlying lava is too
viscous to well up and heal the tears, then the torn-apart portions of skin become rub-
ble fragments (Macdonald, 1953) and the lava flow becomes a'a.
In contrast, at a time of low discharge, flowage of the lava through tubes under the
surface crust is favored. A labyrinth of small single-flow-unit tubes forms, and coales-
cence or enlargement of these causes master tubes to develop. Heat is very efficiently
348
conserved in tubes and lava can travel far from the vent without cooling significantly
(Swanson, 1973). These eruption conditions favor the extreme subdivision of the lava
into flow units, and these units are individually so thin that flowage has almost ceased
by the time that the lava viscosity has reached a value appropriate (in rapidly flowing
lava) to the formation of an a' a surface. A' a does not therefore form and the lava is
predominantly pahoehoe right to its distal limits.
Close monitoring of Kilauea summit tilt during the current (1983-1986) eruptive
activity gives an insight into factors relating to the drastically different eruptive char-
acters. In the time prior to each episode of high discharge and high f ountaining at
Pu'u O'o, the summit reservoir swelled (Wolfe et.al.. 1987). The periodicity of eruptive
episodes is recorded as a saw-tooth pattern on a tilt angle vs. time diagram (cf. Eaton
& Murata, 1960), due to the slow inflation followed by rapid deflation accompanying
eruption. This pattern is interpreted to indicate that magma is stored in a summit
reservoir between eruptive episodes (Wolfe et.al., 1987). In contrast, during eruptive
periods of low or no fountaining (episode 35-B and the current activity), associated with
pahoehoe production, little systematic variation in the tilt record is shown; magma is
not being stored in the summit chamber and output at the vent equals input from the
mantle (Heliker et.al., 1985; Ulrich et.al., 1987).
By analogy, temporal information regarding the 1859 and 1880-81 eruptions of
Mauna Loa allows us to constrain volumes of Mauna Loa summit magma chambers as
well as supply rates to the chamber. The values are approximately 200 x 10^ m^ and 7
mVsec respectively. Dzurisin et.al. (1984, fig. 7) estimate the volume of Kilauea's
chamber as ranging between 50 and 200 x 10^ m^. Estimates of the supply rate to
Kilauea's summit chamber range from 2.7 to 3.5 mVsec (Ulrich et.al., 1986 ; Swanson,
1972; Dzurisin et.al., 1984). Analysis of total area covered by a'a and tube-fed
pahoehoe on Mauna Loa by Lockwood &c lipman (1985) indicates that prior to 750 y ago,
pahoehoe was the dominant eruptive product while since then, a'a has dominated. This
m.ay indicate that since 750 y ago, magma has been stored between eruptions in a sum-
mit chamber.
Consider now the reasons for the diflerences in discharge rate during single erup-
tions as well as for the episodic behavior of Kilauea between 1983 and 1986, that pro-
duced a paired assemblage of a'a and pahoehoe. Three possibilities are considered.
One, the primary control for the episodic activity was partial blockage of the conduit
(between magma chamber and vent) between eruptive episodes. Eventually (by July,
1986) the conduit attained thermal equilibrium, and continuous unimpeded magma
flow became possible. Two, the primary control for the episodic activity was the accu-
mulation of gas bubbles in the magma within the upper part of the magma chamber,
and its depletion lower down. Each eruptive episode released gas-rich magma, and was
then terminated because the underlying gas-poor magma lacked sufficient bouyancy to
rise to the surface. Since then, steady state conditions have been attained, and the
magma emerges continuously with a uniform and moderate content of bubbles. Three,
high eflusion-rate lava is is erupted when strain release on a swelled magma chamber is
the mechanism which drives magma upward to shallow levels, where gas bubble expan-
sion rapidly takes place. The later pahoehoe uses the same conduit to reach the sur-
face at a much lower velocity by means of gas bubble-induced bouyancy.
The relevance of this work to the Martian volcanoes is that two fundamentally
different kinds of long lava flows can be distinguished on Hawaiian volcanoes. The two
kinds may have identical initial viscosities, chemical compositions, flow lengths, and
flow volumes, but their flow mechanisms and thermal energy budgets are radically
different. One travels a distance set by the discharge rate as envisaged by Walker
(1973) and Wadge, (1978), and the other travels a distance set mainly by the eruption
duration and ground slope. In the Mauna Loa lavas, yield strength becom.es an impor-
tant flow-morphology control only in the distal parts of a'a lavas. The occurrence of
349
paired flows on Mauna Loa yields insights into the internal plumbing systems of the vol-
cano, and it is significant that all of the volume of the a' a flow must be stored in a
magma chamber before eruption, while none of the volume of the pahoehoe needs to be
so stored. We are confident that it should be possible to distinguish between the two
kinds of flows on images of Martian volcanoes and hence start to aquire an improved
understanding of these huge structures.
BIBUOGRAPHY
Carr, M.C.. 1981, The Siarface of Mars. Yale University Press, New Haven & London,
232pp.
Dzurisin, D., Koyanagi, R.Y., and English, T.T., 1984, Magma supply and storage at
Kilauea volcano, Hawaii, 1956-1983. J. Volcanol. Geotherm. Res.. 21: pp. 177-206.
Eaton, J.P., and Murata, K.J., 1960, How volcanoes grow. Science, 1^: pp. 925-938.
Emerson, O.H., 1926, The formation of a' a euid pahoehoe. Am. J. ScL, ser. 5, 5: pp. 109-
114.
Heliker, C., Hoflm^ann, J., Rowland, S., Greenland, P., and Reason, M., 1985, Fissiare
activity near Pu'u O'o, Kilauea volcano, Hawaii. Eos, 66: p. 851.
Lockwood, J.P., and Lipman, P.W., 1985, Holocene eruptive history of Mauna Loa vol-
cano, Hawaii. Eos, 66: p. 1132.
Macdonald, G.A-, 1953, Pahoehoe, a' a, and block lava. Am. J. Sci., 251: pp. 169-191.
Malin, M.C., 1980, Lengths of Hawaiian lava flows. Geology, 8: pp. 306-308.
Peterson, D.W., and Tilling, R.L, 1980, Transition of basaltic lava from pahoehoe to a' a,
Kilauea volcano, Hawaii: field observations and key factors. J. Volcanol. Geotherm.
Res., 7: pp. 271-293.
Swanson, D.A., 1972, Magma supply rate at Kilauea volcano, 1952-1971. Science, 175: pp.
169-170.
Swanson, D.A., 1973. Pahoehoe flows from the 1969-71 Mauna Ulu eruption, Kilauea vol-
cano, Hawaii. Geol. Soc. America Bull., 84:: pp. 615-626.
Swanson, D.A., Duflield, W.A., Jackson, D.B., and Peterson, D.W., 1979, Chronological nar-
rative of the 1969-71 Mauna Ulu eruption of Kilauea volcano, Hawaii. U.S. Geol.
Surv. Prof. Pap. 1(S6: 55pp.
Ulrich, G., Heliker, C., Greenland, P., Koyanagi, R., and Griggs, J., 1986, Geology section
inH.V.O. Bulletin for July, 1986.
Ulrich. G.E., Wolfe, E.W., Heliker, C.C, and Neal. C.A. 1987. Pu'u O'o IV: Evolution of a
plumbing system, in "Hawaii symposium on how volcanoes work" abstract volume.
Wadge. G., 1978. Eflfusion and the shape of a' a lava flow fields on Mount Etna, Geology. 6:
pp. 503-506.
Walker. G.P.L., 1973. Lengths of lava flows. PhUos. Trans. Roy Soc. A, 274: pp. 107-118
e. E.W.. Garcia. M.O., Jackson. D.B., Ko
1987, The Pu'u O'o eruption of Kilaue;
1984. U.S. GeoL Soc. Prof. Pap. 1350.
Wolfe. E.W., Garcia. M.O., Jackson. D.B., Koyanagi, R.Y.. Neal, C.A, and Okamura. A.T.,
1987, The Pu'u O'o eruption of Kilauea volcano, episodes 1-20. January 19B3-June
- _ jjj., - "
350
THE 1984 MAUNA LOA ERUPTION AND PLANETARY GEOLOGY
Henry J, Moore^ U.S. Geological Survey, Menlo Park, CA. , 94025
In planetary geology, lava flows on the Moon and Mars are
commonly treated as relatively simple systems. The purpose of
this abstract is to illustrate some of the complexities of
actual lava flows using the main flow system of the 1984 Mauna
Loa eruption. The outline, brief narrative, and results below
are based on a number of sources [1,2,3].
This flow system developed in four distinct stages that
overlap in time; (1) rapid advance of a narrow aa sheet, (2)
development of a channel within the aa sheet that conducted
lava from the vents to the lower reaches, (3) formation of
blockages and obstructions in the channel that produced
overflows, levees, and lava ponds on the aa sheet, and (4)
waning stages during which the lava channels drained and the
distal parts of the flow thinned and spread. Blockages,
obstructions, and overflows progressed upstream with time.
There were also significant variations in the lava flowing in
the channel from the vent toward the toe: (1) densities of
samples from the flowing lava increased from less than 530 to
more than 2,400 kg/m , (2) tempertures of the most fluid lava
decreased from 1140°C to as low as 1086°C, (3) concentrations
and sizes of warm to incandescent objects increased, (4)
apparent viscosity increased dramatically, (5) the rheology
changed, and (6) the volume flow rates decreased.
On the afternoon of March 25, lava was issuing from fissure
vents at the 2850-m elevation that became the principal sources
of lava for the remainder of the eruption. Three southerly
flow lobes fed by the vents stagnated by March 27, but the main
flow was fed at a fairly constant volume-flow rate (560
m /s). Constant-volume flow rates were sustained from March 30
to April 7; after April 7, flow rates declined. The main flow
(flow 1) advanced rapidly as a narrow aa sheet, elliptical in
profile, to the 910-m level, 25 km from the vents. The sheet
flow evolved into (1) a channel zone within the sheet flow
below the vents, (2) a transition zone farther downstream, and
(3) a dispersed-f low zone led by the advancing toe. The lava
channel had developed in the sheet flow by March 29. Also on
March 29, obstructions and blockages near the 1740-m level
caused a channel overflow or breakout that cut off the lava
supply to flow 1 so that it moved about another km in a day.
Lava from this breakout gave rise to flow lA, which advanced
rapidly along a course sub-parallel to and north of flow 1. A
series of overflows that progressed upstream beginning April 3
produced levees and lava ponds that were superposed on the aa
sheet. Blockages and collapse of lava-pond walls gave rise to
surges, ebbs, and small overflows that reduced the lava supply
to flow lA. On April 5, a breakout occurred at about the 1980-
m level and cut off the lava supply to flow lA, which stopped
351
about 27 km from the vents „ The April 5 breakout gave rise to
flow IBj which moved toward the northeast. Repeated channel
blockages and overflows continued progressively upstream from
April 5 to April 8. On April 7, lava production at the vent
began to dwindle; subsequently, the flow system stagnated, vent
activity ceased, and the channels drained.
The appearance of lava during the eruption correlated with
changes in the apparent viscosity of the flow. At the vents
and 3 km away, the flow was composed of sparse cinders and
clinkers in a matrix of molten lava confined in a channel 20 m
wlde» Flow was laminar and steady; velocities were 15 m/s
(vents) and 5.3 m/s (3 km from vents). At 9 km from the vents,
the flow was composed of dark cinders and clinkers, and
incandescent clots in molten lava. At 15 km from the vents,
the flow resembled a slowly moving mass of debris confined in a
rubbly, leveed channel 57 m wide; the flow included warm to
incandescent fragments that were block size and smaller, and
molten lava. Movement occurred by displacement of discrete,
intact units with boundaries that paralleled the crests of the
levees. Flow was laminar but unsteady, with surges and ebbs;
velocities ranged from 0.1 to 0.3 m/s.
Apparent viscosities of the lava were calculated from
observed velocities, assumed densities based on samples from
the flowing lava, and flow dimensions along the main flow. On
a given day, apparent viscosities increased downstream. On
April 2, they were about 10 Pa's at the vents, 10 Pa"s at 3
km from the vents, 10 Pa's at 15 km from the vents, and near
10 Pa's at the toe. These increases in apparent viscosity
were probably related to (1) increases in the concentrations of
solid debris, crystals, and plastic clots, (2) reductions in
gas and bubble contents, (3) decreases in temperatures, and (4)
decreases in stresses and shear rates.
Flow laws probably varied along the length of the flow from
Newtonian, through Bingham, to pseudoplast ic fluids [e.g. 4],
Other f 
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11